Influence of the punisher on the feasibility and sensitivity of a Judgment Bias Task for cattle

Affect influences decision-making: individuals experiencing positive affect interpret ambiguous situations more positively than individuals experiencing negative affect. In animals, decision-making under ambiguity, hence affect, is assessed using judgment bias tasks (JBTs). Since JBT has not yet been designed for cows, we aimed to develop a feasible and sensitive task for cows. Cows were allocated to one of three punishers: “no-reward” (NOTH, n=12), an air puff (AIR, n=12), or an electric shock (ELEC, n=12). Cows were trained to discriminate a positive (P) from a negative (N) cue based on feeder location. Cows learnt to reach (P) within 20s to get concentrates and to not reach (N) during 90s to avoid punisher. Percentage of correct responses to (N) was used to assess JBT feasibility. Data were analyzed using GLMM including the fixed effects of batch and punisher. Cows then faced three ambiguous cues placed between (P) and (N). Latencies to reach these cues were recorded. In a sensitive test, a linear relationship is expected between latencies and distance of the cues from (P) – i.e. latencies decrease when cues are closer to (P). Area between the theoretical line and the experimental profile of responses obtained for each punisher was calculated to assess JBT sensitivity. Data were analyzed using Wilcoxon tests. Percentage of correct responses to (N) was lower for NOTH- compared with AIR- (p=0.032) and ELECcows (p<0.001) (46 ± 4.7%, 66 ± 7.0% and 84.6 ± 2.1%, respectively) during training. Thus, “no-reward” is a less suitable punisher for a feasible JBT since NOTH-cows kept responding to (N). Area between the theoretical and the experimental profiles of responses was smaller for trained AIR-cows than trained ELEC-cows (p<0.001, -0.20 ± 0.186, 0.96 ± 0.140, respectively) indicating that air puff led to a more sensitive test. We therefore advise using air puff as a punisher in JBTs for cows.