Salivary Cortisol, but Not Oxytocin, Varies With Social Challenges in Domestic Pigs: Implications for Measuring Emotions

Abstract

Animals respond to inherently rewarding or punishing stimuli with changes in core affective states, which can be investigated with the aid of appropriate biomarkers. In this study we evaluate salivary cortisol (sCORT) and salivary oxytocin (sOXT) concentrations under baseline conditions and in response to two negatively- and two positively-valenced social challenges in 75 young pigs (Sus scrofa domesticus), housed and tested in eight social groups. We predicted that: (1) Relative to baseline, weaning and brief social isolation would be associated with increases in sCORT, due to psychosocial stress, and reductions in sOXT, due to a lack of opportunities for social support; and (2) Opportunities for social play, and reunions with group members after a separation would be associated with weaker sCORT responses, and increases in sOXT concentrations compared to baseline and to negative social challenges. Testing and sample collection occurred between 28 and 65 days of age and involved a within-subject design, in which every subject was sampled multiple times in neutral (baseline), negative and positive social contexts. We also recorded behavioral data and measured rates of agonism, play and affiliative interactions in the different contexts, prior to saliva sampling. As expected, negative social challenges were associated with robust cortisol responses. Relative to baseline, pigs also had higher sCORT responses to positive social challenges, although these differences were only significant during reunions. Salivary oxytocin concentrations did not differ between the different social conditions, although sOXT was lowest during the brief social isolation. Behavioral analyses confirmed predictions about the expected changes in social interactions in different social contexts, with increases in agonism following weaning, increases in coordinated locomotor play in the play context and high rates of affiliative interactions during reunions. Relative sCORT reactivity to different contexts may reflect the intensity of emotional responses, with greater increases occurring in response to challenges that involve more psychosocial stress. Our results suggest that sOXT is not a reliable indicator of emotional valence in pigs, although more research is needed to characterize sOXT responses to various challenges with and without access to social support.

Keywords: Sus scrofa domesticus; emotions; non-invasive; physiology; play; reunions; social isolation; weaning.

Figure 1

Schematic of the time frame for saliva sampling of each subject. From 28 to 55 days of age, pigs were sampled at similar times of day after exposure to negative, positive or neutral social contexts. From days 60 to 65, each pig was sampled three times on one day before (Base) and during a separation (Sep) from their group members, and after a reunion (Reun).

Figure 2

Log transformed salivary cortisol concentrations following neutral, negative and positively-valenced social contexts. Boxplots indicate medians and interquartile ranges. Sample size for each context is also indicated. Results of the pair-wise post-hoc multiple comparisons (Tukey's HSD) are visualized using a compact letter display, with different letters indicating significant differences. For full model results, see Table 2.

Figure 3

Log transformed salivary oxytocin concentrations following neutral, negative and positively-valenced social contexts. Boxplots indicate medians and interquartile ranges. Sample size for each context is also indicated. For full model results, see Table 2.

Figure 4

Rates of agonistic interactions (A) and locomotor play (B) within social groups during neutral (baseline), negative (weaning) and positive (play) social contexts. The number of social groups sampled in each context is indicated. Boxplots indicate medians and interquartile ranges. Results of the pair-wise post-hoc multiple comparisons (Bonferroni tests) are visualized using a compact letter display, with different letters indicating significant differences.

Figure 5

Proportion of instantaneous scans during which pigs were active in neutral (baseline), negative (weaning) and positive (play) social contexts. The number of social groups sampled in each context is indicated. Results of the pair-wise post-hoc multiple comparisons (Bonferroni tests) are visualized using a compact letter display, with different letters indicating significant differences.

Figure 6

Rates of affiliative and agonistic interactions between each reunited pig and their group members in the 15 min reunion period after a brief social isolation.