The variable vocal behavior of human infants is the scaffolding upon which speech and social interactions develop. It is important to know what factors drive this developmentally critical behavioral output. Using marmoset monkeys as a model system, we first addressed whether the initial conditions for vocal output and its sequential structure are perinatally influenced. Using dizygotic twins and Markov analyses of their vocal sequences, we found that in the first postnatal week, twins had more similar vocal sequences to each other than to their non-twin siblings. Moreover, both twins and their siblings had more vocal sequence similarity with each other than with non-sibling infants. Using electromyography, we then investigated the physiological basis of vocal sequence structure by measuring respiration and arousal levels (via changes in heart rate). We tested the hypothesis that early-life influences on vocal output are via fluctuations of the autonomic nervous system (ANS) mediated by vocal biomechanics. We found that arousal levels fluctuate at ∼0.1 Hz (the Mayer wave) and that this slow oscillation modulates the amplitude of the faster, ∼1.0 Hz respiratory rhythm. The systematic changes in respiratory amplitude result in the different vocalizations that comprise infant vocal sequences. Among twins, the temporal structure of arousal level changes was similar and therefore indicates why their vocal sequences were similar. Our study shows that vocal sequences are tightly linked to respiratory patterns that are modulated by ANS fluctuations and that the temporal structure of ANS fluctuations is perinatally influenced.

A key question for understanding speech evolution is whether or not the vocalizations of our closest living relatives-nonhuman primates-represent the precursors to speech. Some believe that primate vocalizations are not volitional but are instead inextricably linked to internal states like arousal and thus bear little resemblance to human speech. Others disagree and believe that since many primates can use their vocalizations strategically, this demonstrates a degree of voluntary vocal control. In the current study, we present a behavioral paradigm that reliably elicits different types of affiliative vocalizations from marmoset monkeys while measuring their heart rate fluctuations using noninvasive electromyography. By modulating both the physical distance between marmosets and the sensory information available to them, we find that arousal levels are linked, but not inextricably, to vocal production. Different arousal levels are, generally, associated with changes in vocal acoustics and the drive to produce different call types. However, in contexts where marmosets are interacting, the production of these different call types is also affected by extrinsic factors such as the timing of a conspecific's vocalization. These findings suggest that variability in vocal output as a function of context might reflect trade-offs between the drive to perpetuate vocal contact and conserving energy.

Variable motor sequences of animals are often structured and can be described by probabilistic transition rules between action elements. Examples include the songs of many songbird species such as the Bengalese finch, which consist of stereotypical syllables sequenced according to probabilistic rules (song syntax). The neural mechanisms behind such rules are poorly understood. Here, we investigate where the song syntax is encoded in the brain of the Bengalese finch by rapidly and reversibly manipulating the temperature in the song production pathway. Cooling the premotor nucleus HVC (proper name) slows down the song tempo, consistent with the idea that HVC controls moment-to-moment timings of acoustic features in the syllables. More importantly, cooling HVC alters the transition probabilities between syllables. Cooling HVC reduces the number of repetitions of long-repeated syllables and increases the randomness of syllable sequences. In contrast, cooling the downstream motor area RA (robust nucleus of the acropallium), which is critical for singing, does not affect the song syntax. Unilateral cooling of HVC shows that control of syllables is mostly lateralized to the left HVC, whereas transition probabilities between the syllables can be affected by cooling HVC in either hemisphere to varying degrees. These results show that HVC is a key site for encoding song syntax in the Bengalese finch. HVC is thus involved both in encoding timings within syllables and in sequencing probabilistic transitions between syllables. Our finding suggests that probabilistic selections and fine-grained timings of action elements can be integrated within the same neural circuits.

Vocal production is hierarchical in the time domain. These hierarchies build upon biomechanical and neural dynamics across various timescales. We review studies in marmoset monkeys, songbirds and other vertebrates. To organize these data in an accessible and across-species framework, we interpret the different timescales of vocal production as belonging to different levels of an autonomous systems hierarchy. The first level accounts for vocal acoustics produced on short timescales; subsequent levels account for longer timescales of vocal output. The hierarchy of autonomous systems that we put forth accounts for vocal patterning, sequence generation, dyadic interactions and context-dependence by sequentially incorporating central pattern generators, intrinsic drives and sensory signals from the environment. We then show the framework's utility by providing an integrative explanation of infant vocal production learning in which social feedback modulates infant vocal acoustics through the tuning of a drive signal.

Across vertebrates, progressive changes in vocal behavior during postnatal development are typically attributed solely to developing neural circuits. How the changing body influences vocal development remains unknown. Here we show that state changes in the contact vocalizations of infant marmoset monkeys, which transition from noisy, low frequency cries to tonal, higher pitched vocalizations in adults, are caused partially by laryngeal development. Combining analyses of natural vocalizations, motorized excised larynx experiments, tensile material tests and high-speed imaging, we show that vocal state transition occurs via a sound source switch from vocal folds to apical vocal membranes, producing louder vocalizations with higher efficiency. We show with an empirically based model of descending motor control how neural circuits could interact with changing laryngeal dynamics, leading to adaptive vocal development. Our results emphasize the importance of embodied approaches to vocal development, where exploiting biomechanical consequences of changing material properties can simplify motor control, reducing the computational load on the developing brain.