Abstract
Objective and efficient methods for investigating animal's mental states are important for advancing animal welfare research. Behavioural observations and physiological measures provide valuable information but may not fully capture affective states. Judgment Bias Tasks (JBTs) offer a promising method to infer affective states from animals' responses to ambiguous stimuli, as distressed individuals show more pessimistic tendencies. However, the application of JBTs outside research settings may be limited due to time-consuming training and methodological or infrastructural requirements. With the present study, we aimed to test the feasibility of a simple and accessible JBT for non-human primates and to investigate its sensitivity to affective states. Eight brown capuchins (Sapajus apella) and eleven Tonkean macaques (Macaca tonkeana) were tested under positive and negative affective conditions. Individuals were trained to discriminate between never rewarded conditioned stimuli and always rewarded ones. They were then presented with ambiguous probes they had never seen before to analyse their responses to ambiguity. The discrimination training required, on average, about 12 sessions lasting a few minutes each to successfully train every individual animal. Both species significantly discriminated between conditioned stimuli and ambiguous probes. Additionally, macaques showed a greater tendency to interact with the never rewarded condition stimuli and the probes under positive conditions during early sessions. In capuchins, however, the affective conditions presented had no significant effect, although the descriptive analysis tended to show that the latency to investigate the probes was greater in the negative condition. This study also highlights key methodological considerations, such as a decrease in ambiguity across sessions. Overall, the present results suggest that, with a few improvements, this method could be effectively implemented to assess affective states of primates under human care, supporting this tool as a valuable candidate for welfare assessment.