How does the previous environment affect pigs’ responses when switching to a barren or enriched environment as assessed by Qualitative Behaviour Assessment?

Abstract

Intensive pig farming systems are often characterised by barren and monotonous environments, which can impair pig welfare. While enriching the environment enhances pig welfare, less is known about the implications of changes in housing conditions regularly happening on commercial farms. This study investigated the effect of a previously enriched-varying (E) or barren-monotonous (B) environment on pigs’ behavioural expressions when exposed to a switch of environments from E to B or from B to E as assessed by Qualitative Behaviour Assessment (QBA). QBA is based on the assumption that human observers can integrate subtle details in animals' body language to reflect their affective experience by using descriptors like 'lively' or 'apathetic'. Eight groups of pigs underwent a switch of housing conditions from E to B or from B to E (E*B, B*E), while another eight groups moved to different pens, but with the same conditions (E*E, B*B). Pigs’ behaviour was video recorded 24 h before (PRE) and 24 h after the switch (POST). Using Free Choice Profiling, 14 observers blinded to the PRE-switch housing conditions were asked to score pigs’ behavioural expressions on 64 video clips. Two main dimensions (D1, D2) were identified, which accounted for 65.4 % of the variation. Terms describing dimension 1 (D1) ranged from ‘active/engaged/explorative’ to ‘tired/bored’, while dimension 2 (D2) was characterised as ranging from ‘bored/restless/aggressive’ to ‘relaxed/content/calm’. Against expectation, there was no statistical support for an interaction between PRE-switch and POST-switch condition on pigs’ behavioural expression POST-switch (D1: F_1,11 = 0.03, p = 0.86; D2: F_1,11 = 0.57, p = 0.47). We cannot disentangle if the PRE-switch conditions did not affect pigs’ behavioural expression POST-switch or if observers did not pick up potential changes, e.g. because the contrasting POST-switch conditions may have masked subtle effects from the PRE-switch conditions. However, even though not statistically supported, the direction of effects was according to hypotheses with E*B groups being more uniformly assessed as reflecting a negative affective state (tired/bored) compared to B*B, while B*E groups were more consistently judged as reflecting a positive affective state (active/engaged/explorative) compared to E*E. To test if this direction of effect is robust, we propose to show video clips of pigs in the same condition POST-switch (either E or B, not both) with different experiences in PRE-switch conditions in future studies. Moreover, future research should provide insight into what changes in housing conditions mean for pigs’ affective states in the longer term.