Understanding animal emotions is a key to unlocking methods for improving animal welfare. Currently there are no ‘benchmarks’ or any scientific assessments available for measuring and quantifying the emotional responses of farm animals. Using sensors to collect biometric data as a means of measuring animal emotions is a topic of growing interest in agricultural technology. Here we reviewed several aspects of the use of sensor-based approaches in monitoring animal emotions, beginning with an introduction on animal emotions. Then we reviewed some of the available technological systems for analyzing animal emotions. These systems include a variety of sensors, the algorithms used to process biometric data taken from these sensors, facial expression, and sound analysis. We conclude that a single emotional expression measurement based on either the facial feature of animals or the physiological functions cannot show accurately the farm animal’s emotional changes, and hence compound expression recognition measurement is required. We propose some novel ways to combine sensor technologies through sensor fusion into efficient systems for monitoring and measuring the animals’ compound expression of emotions. Finally, we explore future perspectives in the field, including challenges and opportunities.

Animal welfare is an important field of study due to animal sentience, yet there is to date no consensus on the definition of animal welfare. There have been four key developments in the field of animal welfare science since its birth: the theoretical and empirical study of affective states, and hence our understanding thereof, has increased; there has been a shift from a primary focus on unpleasant experiences towards an inclusion of pleasant experiences; there has been an increasing mention and investigation of the notion of cumulation of experiences in time, and with this, the importance of the time component of both affective states and animal welfare has come forward. Following others, we define welfare as a balance or cumulation of pleasant and unpleasant experiences over time. The time period of welfare depends on when welfare considerations are necessary, and may range from the duration of single and relatively short-term experiences to the entire life of an animal. We further propose that animal welfare conceptualised in this way can be assessed at three levels: level 1 represents the assessment of the environment and 'internal factors' such as health and personality, which interact in their impact on the affective experiences of animals; level 2 represents the assessment of affective states; and level 3 represents the assessment of the balance or cumulation of these affective states in time. The advancement of research necessitates studies to be more or less comparable, and this would be facilitated by researchers mentioning which concept of animal welfare they are basing their work on, at which level of assessment they are working, which assumptions they might be drawing from to infer welfare and which time period of interest they are focusing on, even if this is not mirrored by the timing of the assessment in practice. Assessment at levels 2 and 3 still needs much study, at both the theoretical and empirical levels, including agreements on validation tools.

We studied emotional contagion, a simple form of empathy, and the role of oxytocin herein in pigs. Two training pigs per pen (n = 16 pens) were subjected to a positive treatment (pairwise access to a large compartment filled with peat, straw and some chocolate raisins) and a negative treatment (social isolation in a small compartment) in a test room using a within-subjects design. Thereafter, two naive pen mates joined the training pigs in the test room, but were not given access to the treatments. This allowed testing for emotional contagion. Subsequently, the naive pigs, serving as their own controls, were given 24 IU of oxytocin or a placebo intranasally 30 min before accompanying the training pigs, which were exposed to either the negative or positive treatment, to the test room. Behavioral differences found between the positive and negative treatments (e.g., play and “tail wagging” vs. standing alert, urinating, defecating and ears backward) show that the treatments induced a positive and negative emotional state in the training pigs, respectively. Changes in behaviors of the training pigs with and without naive pigs present (e.g., in ears backwards) and of the naive pigs with and without training pigs present (e.g., in standing alert) indicated that emotional contagion occurred, especially during the negative treatment. Oxytocin did not seem to affect the behavior of the treated naive pigs, but did affect behaviors (e.g., defecating) of the training pigs which had not received oxytocin. This suggests a role for oxytocin in pig communication, which merits further research.

Pigs living in intensive husbandry systems may experience both acute and chronic stress through standard management procedures and limitations in their physical and social environment, which may have implications for their immune status. Here, the effect of a new breeding method where pigs were selected on their heritable influence on their pen mates' growth, and environmental enrichment on the immune status of pigs was investigated. Hereto, 240 pigs with a relatively positive genetic effect on the growth of their pen mates (+SBV) and 240 pigs with a relatively negative genetic effect on the growth of their pen mates (−SBV) were housed in barren or straw-enriched pens from 4 to 23 weeks of age (n = 80 pens in total). A blood sample was taken from the pigs before, three days after a 24 h regrouping test, and at week 22. In addition, effects of coping style, as assessed in a backtest, and gender were also investigated. Mainly, +SBV were found to have lower leukocyte, lymphocyte and haptoglobin concentrations than -SBV pigs. Enriched housed pigs had a lower neutrophil to lymphocyte (N:L) ratio and lower haptoglobin concentrations, but had higher antibody titers specific for Keyhole Limpet Hemocyanin (KLH) than barren housed pigs. No interactions were found between SBV class and housing. Furthermore, pigs with a proactive coping style had higher alternative complement activity and, in the enriched pens, higher antibody titers specific for KLH than pigs with a reactive coping style. Lastly, females tended to have lower leukocyte, but higher haptoglobin concentrations than castrated males. Overall, these results suggest that +SBV pigs and enriched housed pigs were less affected by stress than -SBV and barren housed pigs, respectively. Moreover, immune activation might be differently organized in individuals with different coping styles and to a lesser extent in individuals of opposite genders.