Welfare includes both physical health and emotional well-being, and cognitive biases have recently been proposed to reflect animals` underlying mood state. Several test paradigms to assess judgement biases have been developed for various species. In horses, spatial judgement bias tests have been used but it is unclear whether the latency measurements provide valid measures of emotional valence. The aim of this study was to compare two test paradigms based on choices, a Go/No-go paradigm (GNgP) and an Active Choice paradigm (ACP), in terms of time needed for training and variation in test outcomes. Five stallions and five mares were randomly assigned to the two paradigms and trained for 10 trials per day to discriminate between a low and a high tone. In the GNgP, horses learned to approach a bucket for a food reward when presented with one tone, and to stay in the start box when presented with the other. In the ACP, two buckets were presented and the horses were trained to approach the bucket on one side in response to one tone to obtain a big reward and to approach the bucket on the other side in response to the other tone to obtain a small reward. After reaching a predetermined learning criterion, horses were tested for four days with three intermediate tone-trials interspersed between 10 high and low tone-trials. All five GNgP horses but only one ACP horse reached criterion in an average of 38 training sessions (ranging from 25 to 52 sessions). Training was stopped after 40 sessions in the remaining ACP horses whose performance did not improve. During testing, the GNgP horses showed an intermediate response with a slightly positive bias (63%) to the three intermediate tones without discriminating between them. Moreover, individual horses responded differently to the intermediate tones. Thus, horses can be successfully trained on a GNgP based on auditory cues, while a comparable ACP may be too difficult for them to learn. However, further refinements will be needed to render this GNgP a valuable tool for assessing emotional valence in horses. LP: Cognitive bias tests provide a useful tool for the assessment of horses` emotional states and therefore horse welfare. The test is based on the assumption that emotions influence judgement and that we can in turn draw conclusions from these judgements about a horse`s underlying emotional state.
© Copyright 2014 DCA Report; Nr. 044. Veröffentlicht von University of Aaarhus. Alle Rechte vorbehalten.