Assessing animal affect: an automated and self-initiated judgement bias task based on natural investigative behaviour

Samantha Jones,Vikki Neville,Peter Dayan,Emma S. J. Robinson,Laura Higgs,Elizabeth S. Paul,Michael Mendl

doi:10.1038/s41598-018-30571-x

Abstract

Scientific methods for assessing animal affect, especially affective valence (positivity or negativity), allow us to evaluate animal welfare and the effectiveness of 3Rs Refinements designed to improve wellbeing. Judgement bias tasks measure valence; however, task-training may be lengthy and/or require significant time from researchers. Here we develop an automated and self-initiated judgement bias task for rats which capitalises on their natural investigative behaviour. Rats insert their noses into a food trough to start trials. They then hear a tone and learn either to stay for 2 s to receive a food reward or to withdraw promptly to avoid an air-puff. Which contingency applies is signalled by two different tones. Judgement bias is measured by responses to intermediate ambiguous tones. In two experiments we show that rats learn the task in fewer sessions than other automated variants, generalise responses across ambiguous tones as expected, self-initiate 4–5 trials/min, and can be tested repeatedly. Affect manipulations generate main effect trends in the predicted directions, although not localised to ambiguous tones, so further construct validation is required. We also find that tone-reinforcer pairings and reinforcement or non-reinforcement of ambiguous trials can affect responses to ambiguity. This translatable task should facilitate more widespread uptake of judgement bias testing.

Highlights

Scientific methods for assessing animal affect, especially affective valence, allow us to evaluate animal welfare and the effectiveness of 3Rs Refinements designed to improve wellbeing
There was no significant effect of positive tone frequency (2 vs 8 kHz) on the number of sessions to criterion during the discrimination training (DT) phase (Log-rank chi-square = 0.10, df = 1, p = 0.919), the fully- or partially-reinforced discrimination training (FPRDT) phase (Log-rank chi-square = 0.765, df = 1, p = 0.382), or across both phases (Logrank chi-square = 0.239, df = 1, p = 0.625)
Mean percentage correct responses to positive (‘stay’) and negative (‘leave’) tones and mean number of trials initiated per minute are shown in Fig. 2b,c for those sessions before criterion could be achieved in which all rats participated, and for the final two training sessions before judgement bias testing

Summary

Introduction

Scientific methods for assessing animal affect, especially affective valence (positivity or negativity), allow us to evaluate animal welfare and the effectiveness of 3Rs Refinements designed to improve wellbeing. We develop an automated and self-initiated judgement bias task for rats which capitalises on their natural investigative behaviour Rats insert their noses into a food trough to start trials. We find that tone-reinforcer pairings and reinforcement or non-reinforcement of ambiguous trials can affect responses to ambiguity This translatable task should facilitate more widespread uptake of judgement bias testing. One promising approach designed to assess affective valence involves testing an animal’s decision-making under ambiguity (‘judgement bias’)[9,10] This is based on empirical investigations showing that people in negative affective states are more likely to make negative judgements about the meaning or outcome of ambiguous stimuli or situations[11,12]. There are some null or apparently opposite findings in the more than 100 experimental studies in animal welfare science, neuroscience and psychopharmacology that have used variants of this generic method in mammals, birds and insects, the majority of published findings support the a priori prediction[10,15]

Objectives

Methods

Results

Conclusion