Looking out for danger: Theoretical and empirical issues in translating human attention bias tasks to assess animal affective states

Understanding animal emotional (affective) state is highly relevant to various disciplines (e.g., animal welfare, neuroscience, comparative psychology), and has been significantly advanced by translating affect-induced cognitive bias paradigms rooted in human psychology to non-human animal studies. Attention bias (i.e., preferential attention allocation, AB) tests are increasingly used as more practical substitutes to commonly used judgement bias tests. Yet, evidence that AB reflects affective valence in animals is still limited. We review in-depth the concept of attention and AB described in humans and discuss utilising human-derived AB paradigms for measuring animal affective states. We describe key concepts and functions of attention in humans, before concentrating on the relationship between AB to threat detection and human anxiety. We critically review animal AB studies, discuss methodological discrepancies in such studies, and highlight the need for further experimental refinements. This includes identifying appropriate species-specific test designs and stimuli, modes of presentation (e.g., real-life vs. artificial stimuli), and consideration of subject-related factors (e.g., personality, age). We conclude that experimental limitations currently hamper the validity of AB as a proxy of animal affect and hope that the knowledge gaps highlighted in our review will encourage further research.