Goats who stare at wolves - identifying natural response stimuli for an affect-driven attention bias test in small ruminants

Reliable and non-invasive assessment of affective states is crucial in animal welfare research. Linking affective states to cognitive processes, such as cognitive biases, provides a promising approach. An affect-driven attention bias - the tendency to focus more on threatening than on neutral or positive stimuli when in a negative affective state - has been observed in humans and non-human animal species. However, implementation is often hampered because ecologically relevant test setups need to be adapted to species-specific differences in the perception and processing of information from different modalities. As an initial step towards developing and validating an affect-driven attention bias test for goats, we aimed to identify visual stimuli perceived as potentially threatening using a looking-time paradigm. In a within-subject design, 30 adult female dwarf goats (Capra hircus) were presented with photographs of 12 animal species (six natural predators and six non-predators) from three taxa (mammals, reptiles, birds). Each image was shown for 10 seconds on one of two video screens, while the opposite screen remained white. Each goat completed four sessions (one per day) of six trials each: two sessions showing full-body and two sessions showing face/head-only images, resulting in 24 trials per subject. Analysis of relative looking duration revealed a significant interaction between animal taxon and predator category. Goats looked longer at predatory than non-predatory reptiles, while the opposite pattern was found for mammals. No difference was observed for birds. The presented body part (full body vs. face/head only) did not influence looking behaviour. These results suggest that photographs of predatory reptiles, particularly snakes, might be perceived as potentially threatening by goats, indicating their suitability as negative stimuli in future affect-driven attention bias tests. Further validation is needed to confirm their negative valence.