The effects of housing conditions on judgement bias in Japanese quail

Abstract

The interaction between cognition and emotions represents previously unforeseen source of information for the animal welfare science. The prototype study of Harding et al. (2004) draw the attention of many researchers to the phenomenon of cognitive bias in animals. Their original approach using operant discrimination training as the basis of the judgement bias test, however, was replaced in many later studies by methods requiring less training, such as the spatial judgement task. The aim of this study was to validate the operant conditioning based cognitive bias test for Japanese quail and investigate the effect of housing conditions (wire cage vs. deep litter pen) on the affective states using this test. Housing conditions were chosen since they may have a wide range of influences on poultry, ranging from behaviour to brain morphology. Quail hens were first trained in an operant discrimination task (contrasting shades of grey) to respond by pecking at the positive cue associated with reward and to refrain from pecking at the negative cue to avoid punishment. The significant differences in the proportion of responses to positive and negative cue were observed very soon (session 1–3) and the quails reached the discrimination criterion (successive discrimination on 3 consecutive days) in 7–9 sessions. In the judgement bias tests the quails were in addition to pretrained cues exposed to three ambiguous cues (shades of grey intermediate between the positive and negative cue). We hypothesised that the quail hens kept in the deep litter pens will be more ‘optimistic’ in ambiguous situations than the quails kept in the wire cages. In two out of four experiments there were no significant effects of housing on the proportion of responses to the ambiguous stimuli. In one experiment deep litter housed birds responded more to the ambiguous stimuli than the caged ones, but after the exchange of environments the same birds housed in cages still responded more to the ambiguous cue. In the last experiment not only the responses to the ambiguous stimuli were lower in deep litter housed quails, but also the responses to the positive cues. Based on these results we have concluded that either the differences between the housings were not large enough, or the length of the housing was not long enough, to induce stable behavioural differences detectable by the operant cognitive bias tests. Using the given judgement bias test design we have failed to validate it as a tool for the assessment of changes in affective states induced by housing conditions in Japanese quail. Nevertheless, further refinement of the test design especially in relation to possible loss of ambiguity problem is necessary.