• attentional biases
• empathy
• interspecific contagion
• motor mimicry
• Old World monkeys
• yawn contagion

Yawn contagion (YC), highly familiar in our species, has been studied with experimental and/or naturalistic studies in several primate and mammal species and, differently from spontaneous yawning, seems restricted to highly social ones. Neurobiological, psychological, and ethological findings indicate the phenomenon as an expression of the basal layer of empathy. Specifically, YC has been accounted to represent the basal layer in the Russian-doll model for empathy. Supporting evidence comes, for instance, from the fact that contagion has been found to be biased towards affiliated/kin individuals. Indeed, according to the Emotional Bias Hypothesis (EBH), different degrees of emotional closeness with a yawner result in differently contagious stimuli. On the other hand, the Attentional Bias Hypothesis (ABH) states that familiar yawns are more contagious simply because more attention is paid to socially close or relevant subjects. The species considered in this study, the red-capped mangabey (Cercocebus torquatus), is a semi-terrestrial endangered Cercopithecid species native to the West African coastal rainforest regions. Cercocebus spp. live in multi-male multi-female groups, variable in size from 10-20 up to more than 100 individuals. They show high cohesion level between members of medium-sized parties, frequent peaceful and agonistic interactions, steep hierarchies but relatively dynamic dominance relationships, and a rich repertoire of communicative signals, including facial expressions. Here, we used an experimental approach in 17 captive red-capped mangabeys hosted at the Station Biologique de Paimpont (University of Rennes 1) to test the contagiousness of video stimuli from three different species (conspecifics, humans, hamadryas baboons). For conspecifics and humans, it was also possible to have stimuli with two different degrees of familiarity (i.e., familiar vs unfamiliar) created ad hoc for the tested subjects. For each of the 5 types of stimuli (i.e., familiar conspecifics, unfamiliar conspecifics, familiar humans, unfamiliar humans, h. baboons) we had Test and Control conditions (yawning individuals vs non-yawning individuals in the same context), and thus each of the 17 subjects underwent 10 sessions. We used Generalized Linear Mixed Models (GLMMs) to investigate what is affecting the attention levels to the screen (measured as rates of frontal looking to the screen), the likelihood of yawn responses, and anxiety levels (using as proxy data the rate of self-directed behaviours, SDBs) during the experimental sessions. Both intra- and interspecific YC were detected, and the likelihood of yawning was not influenced neither by the species of the yawner, nor by attentiveness to the screen. Conversely, the degree of familiarity with the yawner significantly favoured the contagion: familiar yawns (from both conspecifics and humans) were more contagious than unfamiliar ones. Attention was biased towards conspecifics (probably holding greater socio-ecological relevance compared to other species) and, importantly, unfamiliar humans and mangabeys elicited greater attention than the familiar counterparts, possibly for the potential adaptive implications involved when facing actual live novel stimuli (e.g., unknow subjects). The rates SDBs (i.e., anxiety levels) were comparably low during the different experimental sessions and were not influenced by the kind of stimuli to which they were exposed to; this suggests that anxiety levels did not act as a confounding factor for the detection of YC. Taken together, our results indicate that even though videos with familiar subjects were looked at for shorter, they were triggering yawn contagion more often. Our results thus support the Emotional Bias Hypothesis, since a different degree of familiarity (i.e., of emotional sharing and attachment) resulted in variability in the stimuli contagiousness, whereas they do not give support the Attentional Bias Hypothesis, further suggesting that selective attention may not be involved in susceptibility to YC.