Page 71 - Emotions through the eyes of our closest living relatives- Exploring attentional and behavioral mechanisms

Attention towards humans’ and bonobos’ emotion
additional emotional category, i.e., “self-scratching”) included 60 trials. The human scene dot-probe task included two threat categories, namely distress and aggression. For children there was an equivalent category of pictures for aggression showing bullying behavior.
Statistical analyses
All data were analyzed using generalized mixed modeling with experimental trials
nested within subjects (SPSS; version 25). A random intercept of participant was 3 included in all analyses. Significant interactions and main effects of Emotion Category
were followed-up using t-tests. In order to reduce type I errors, the p-value was set to
0.01. Follow-up simple contrasts are Bonferroni-corrected. Only the significant effects
of major theoretical interest are visualized. We start with reporting main effects,
followed by two-way interactions, followed by higher order interactions, in the order
of significance (most significant effects first).
Rating scales
The dependent variables in this emotion perception task were Emotional Valence and Intensity (1–7 scaled ratings). We used a linear distribution function to model the data, as they were normally distributed. Three separate models were conducted with minor differences. All models included the fixed factors of Age Group (Child/ Adult), Gender (Male/Female), Emotion Category (Sex [adults only]/Groom/Play/ Self-scratch/Distress/Aggression [adults and human scenes only]/Yawn/Neutral) and Species Scene (Human/Bonobo). In the first model, we examined the effect of age, gender, expressor species, and emotional display on intensity ratings. Therefore, we omitted the categories Sex and Aggression since a) children did not see these images and b) the aggression category was not included in the bonobo scenes. Therefore, this model included scenes depicting grooming, yawns, play, distress, and self- scratching. In the second model, we zoomed in on human adults and therefore added the emotion category Sex. In the third model, we included only human adults and human scenes, and added the emotion category Aggression.
Attentional bias
The reaction times were filtered with the following procedure. First, age categories were created using 5-year bins (e.g., 0–5 years old; 5–10 years old) and trials with extremely fast (<250 ms) or extremely slow responses (>5000 ms) were excluded. Next, the trials exceeding 2.5 mean absolute deviations (MAD) from the mean per
69