Assessing the Validity of Abnormal Repetitive Behaviours as Indicators of Poor Animal Welfare: A Narrative Review

Abnormal repetitive behaviours ('ARBs') in captive animals are a heterogeneous group of troubling activities (e.g. stereotypic pacing by Carnivora; feather-plucking by poultry). To assess and improve their construct validity as welfare indicators, we reviewed their responsiveness to mood-improving drugs; links with imprisonment, stress and self-reported poor well-being in humans; and in animals, the impact of welfare-compromising (e.g. aversive) treatments. Considerable evidence links ARB development with negative moods or mood disorders, and early/prolonged/recurrent negative experiences (potentially via dose-response-like effects). Findings also indicate effects of cumulative affective experience ('cumulative stress'). Furthermore, if ARBs transiently help subjects cope, such effects are only partial. Therefore, whenever husbandry or housing causes ARB-prone phenotypes, negative affect can reliably be inferred (with more severe ARBs indicating poorer welfare). However, ARBs are rather prone to false nulls as welfare indicators: prolonged negative affect does not always cause the emergence or increase of ARBs, primarily due to threshold effects, ceiling effects, and inactivity being an alternate response. Furthermore, in ARB-prone subjects, the onset/offset of bouts appears not to reliably track moment-by-moment levels of negative emotion. Additionally, because variation in activity, behavioural flexibility and stress-response style are potential confounds, ARBs are not advised for comparing welfare across individuals, strains, species or prenatal treatments. Overall, ARBs have strong construct validity as indicators of negative moods/mood disorders; and our additional rules-of-thumb should further refine their accuracy. Future research should investigate underlying mechanisms (e.g. those suggested by human and biomedical findings), especially to clarify the boundaries and biological sub-types of ARBs.