Page 77 - Emotions through the eyes of our closest living relatives- Exploring attentional and behavioral mechanisms

Attention towards humans’ and bonobos’ emotion
An interaction between Emotion Category, Species and Sex (F(7, 10.13) = 4.25, p < .001)
showed a significant gender difference in several emotion categories. Specifically,
human distress, human and bonobo yawn images as well as neutral, grooming and
playful bonobo scenes were perceived more intensely by women compared to men
(human distress: t(10.13) = 3.74, p < .001; human yawn: t(10.13) = 2.84, p < .001;
bonobo yawn: t(10.13) = 3.75, p < .001; neutral bonobo scenes: t(10.13) = 4.52, p <
.001; bonobo grooming t(10.13) = 6.75, p < .001) and bonobo play t(10.13) = 3.71, p =
.002). 3
In a third analysis in human adults, we zoomed in on the human scenes specifically, and added the category Aggression. The results showed a main effect of Emotion Category (F(1, 2.62) = 141.24, p < .001) and an interaction between Emotion Category and Sex (F(1, 2.62) = 2.22, p = .008). Of all categories, Aggression was perceived most intensely and Neutral least. Further, the only gender difference that remained significant after having excluded the bonobo scenes from the analysis was the category Distress, with females rating these scenes as more intense than males (t(2.62) = 2.78, p = .006).
Attentional bias toward human and bonobo emotional scenes
Reaction times were analyzed in a generalized mixed model with the fixed factors Age Group, Gender, Species Scene, Emotion Category, Congruency (i.e., the probe appearing behind the emotional scene or not), their interactions, and in addition Dominant Hand Distance (i.e., location of probe in relation to handedness of participant). Again by using a backward elimination method, we came to the final, most parsimonious and best-fitting model that included a subset of these factors. For brevity, only significant effects that include the factor Congruency are described in the text.
In a first analysis where the categories Sex and Aggression were excluded, participants (adults and children), showed a main effect of Congruency (F(1, 29.07) = 6.78, p = .009), demonstrating that reaction times were faster when the probe replaced an emotional compared to a neutral scene, indicating heightened attention for the emotional category. An interaction between Congruency and Species (F(1, 29.07) = 6.84, p = .002), showed that the Congruency effect was significant for human scenes (t(29.07) = 3.76, p < .001), but not for bonobo scenes (p = .698, see Figure 2E). There were no gender differences observed (ps = .441 in a pre-final model) and Age Group did not significantly modulate the congruency effect either (p = .053, trending toward a larger effect in adults).
75