Musicality is the predisposition to process and produce music (Honing et al., 2015). Several hypotheses have been proposed to explain the evolutionary origins of musicality such as promoting social bonding (Music for Social bonding Hypothesis, Dunbar, 2023; Savage et al., 2021), credible signaling (Mehr et al., 2021), facilitating individual identification (Patel and von Rueden, 2021) or coordinating activities (Mogan et al., 2017). Most of these theories have in common a strong link between musicality and sociality in humans with a striking illustration in human children, where synchronizing to a common rhythm promotes prosocial behaviour and cooperation (Rabinowitch and Meltzoff, 2017; Wan and Zhu, 2021). Human musicality also involves a strong cultural component, which underlies its multiple forms in various societies. According to Podlipniak (2023), the evolution of human musicality is supported by cognitive plasticity at the basis of this cultural transmission (see also Killin, 2018; Patel 2021; Savage et al., 2021). Indeed, human musicality frequently involves coordinating to new rhythms, suggesting an open-ended ability to learn new behaviours to adapt to new stimuli.
In addition to cognitive plasticity, several components of musicality are commonly found throughout human musical cultures and occasionally in non-human species (Bouwer et al., 2021; Honing et al., 2015). Beat induction, which allows us to perceive a regular pulse (Patel, 2009), together with relative pitch, or tonal encoding of pitch, are among these basic components. Investigating these basic components in other social species could lead to a better understanding of the evolutionary origins of musicality. Beat induction affords entrainment, i.e., the ability to synchronise behaviour to external rhythms, from tapping in time to a rhythmic stimulus to elaborate multi-participant dances and songs. Entrainment and other rhythm production may be either reactive, i.e. reacting only to past rhythmic events, or predictive, i.e. predicting and matching future rhythmic events, which can imply different internal representations of rhythm and thus indicate different cognitive processes (Bouwer et al., 2021; Kotz et al., 2018; Ravignani et al., 2013).
Recent literature suggests that entrainment originates in neural oscillations, which may themselves be based on basic neural circuitry conserved throughout animal evolution (Rouse et al., 2016; Wilson and Cook, 2016). Yet, evidence for entrainment has remained scarce among mammals and birds (Wilson and Cook, 2016). Most studies have looked for entrainment in non-human primates (e.g. chimpanzees: Takeya et al., 2017; Hattori and Tomonoga 2021: bonobos: Large et al., 2015; macaques: Nagasaka et al., 2013; Selezneva et al., 2013; Takeya et al., 2018; Zarco et al., 2009), with comparatively few studies in other taxa (sea lion: Cook et al., 2013; Rouse et al., 2016; rat: Katsu et al., 2021; budgerigar: Hasegawa et al., 2011; cockatiel: Patel et al., 2009). Moreover, the vast majority of studies on entrainment has relied on prior conditioning for the specific task rather than on spontaneous entrainment. Non-human primates, despite their close phylogenetic relationship to humans and having some rhythmic abilities, do not appear to share the human capacity for entrainment. For instance, chimpanzees react to rhythms but perform poorly at actually keeping the beat (Hattori and Tomonaga, 2020) and show only loosely-coordinated displays, whether through vocalisations (Ghiglieri, 1984; Merker et al., 2009) or body movements (Hattori and Tomonaga, 2020; Lameira et al., 2019; Schweinfurth et al., 2022). Great apes are capable of social learning and thus of cultural transmission of behaviours (e.g. tool use: van Schaik et al., 1990:, context-specific gestures: Malherbe et al., 2025), including drumming (Eleuteri et al., 2025; van Loon et al., 2025), but unlike humans, they are not open-ended vocal learners, which may partly explain their poorer skills for sound entrainment. By contrast, some species of songbirds are open-ended vocal learners, and may thus be better candidates to study entrainment abilities (Rothenberg, 2014; Snyder and Creanza, 2021).
Several parallels have been drawn between birdsong and human music or language. First, species with demonstrated entrainment tend to also be social species, in line with a purported link between sociality and musicality (Podlipniak, 2023; Ravignani et al., 2013; Savage et al., 2021; Wilson and Cook, 2016). Second, birdsong and human music share similarities in temporal organisation (Sainburg et al., 2019; Snyder and Creanza, 2021), and in the neural pathways activated in the respective species (i.e. humans for music, birds for birdsong) when listened to (Earp and Maney, 2012) or performed as a group (Stevenson et al., 2020; Riters et al., 2019). More precisely, when exposed to a rhythm the auditory sensory-motor areas of the human brain (Grahn and Rowe., 2009; Zatorre et al., 2007) and the songbird song circuit (Lampen et al., 2019) both show higher activity when synchronising to a rhythm, even at times when no outward behaviour occurs (i.e. no body movement, no vocal output). Some parrot species, including cockatoos (Patel et al., 2009) and budgerigars (Hasegawa et al., 2011; Seki et al., 2019), are able to extract the beat from music and adapt their movements to it (Fitch, 2009; Patel et al., 2009). Amongst all songbirds, parrots and corvids could be good models to test vocal entrainment in the avian clade. Both groups of species can imitate sounds (including human or other species voices), an ability that could be necessary for rhythmic learning and entrainment abilities (Dooling et al., 2002; Schachner et al., 2009; Schachner, 2010). A few studies have evidenced cultural transmission of calls and thus open-ended vocal learning in various corvids (Brown, 1985; Enggist-Dueblin and Pfister, 2002; Kondo, 2021). Corvids can also produce duetting behaviour demonstrating their ability to coordinate vocal production with conspecifics (Kondo et al., 2010; Seed et al., 2007). Corvids can also learn to produce particular vocalisations in response to particular stimuli, suggesting volitional control of their vocal output (Brecht et al., 2019; Liao et al., 2024; Tomasek et al., 2023). Some corvid species also produce undirected songs, i.e. a series of varied vocalisations (variously described as squawks, sneezes, snores or cackles) that may exhibit high structural variation between occurrences (Brown et al., 1985; Brown and Farabaugh, 1997; Coombs, 1960; Tomasek et al., 2023; Martin et al., 2024). Among corvids, rooks are well-known for their social cognition skills and complex social lives (Clayton and Emery 2004; Coombs, 1960). A previous study evidenced the ability of rooks to spontaneously sing in response to external stimuli, not only starting but also stopping vocal production within a few seconds of the stimulus starting and stopping, respectively (Tomasek et al., 2023). This result demonstrates that rooks are capable of coordinating their vocalisations to an external sound without conditioning or reinforcement; it does not give any indication of their capacity to entrain to a rhythm, also because of a timescale longer than that of music.
The present study tested whether rooks can vocally entrain to a rhythmic, biologically non-relevant stimulus without prior conditioning. Birds were individually exposed to stimuli with different rhythms, varying in both tempo and metrical structure. Whenever this exposure led to the bird singing, we recorded these songs and evaluated whether birds adjusted their songs to the tempo or to the metrical structure. We predicted that if rooks have the ability to entrain spontaneously to the stimulus, they should adjust their song to the tempos and structures of the stimulus they hear. We additionally investigated whether any adjustment could be more compatible with reactive (if adjustment matches to the preceding stimulus onset) or predictive (if adjustment matches the closest stimulus onset, whether before or after) timing.