Looking Into Their Eyes: Using Facial Expressions To Guide Welfare

As most animal advocates will attest to, recognizing, understanding, and managing pain is critically important for animal welfare. However, many of us are frustrated by the current level of understanding of this “aversive feeling” in animals. Surely, the issue is complex, made especially enigmatic by the subjectivity of pain response. Facial expression scales have been used in recognizing and evaluating pain in humans for a long time. In this study, a group of researchers from the U.K., Italy, and Germany joined forces to review existing literature on developing facial expression scales suitable for understanding pain in non-human animals.

It is well-known that pain negatively influences natural animal behaviour: it reduces play, grooming, eating, and disrupts sleep. Unfortunately, several routine husbandry procedures in animal farming such as castration and de-horning can result in pain when carried out with inadequate pain relief. This potentially leads to millions of animals experiencing pain that could be avoided worldwide. Experimental procedures in laboratories, accidental injuries, diseases, or elective surgeries may also result in pain. Unmitigated pain can result in pathological changes in physiology and behaviour, decreasing the validity of scientific studies involving animals even further beyond their typically poor translatability to human beings. This is besides the obvious fact that the animals may end up suffering from chronic pain as a result.

If we do not know how to identify pain and judge its severity, we simply cannot expect to be able to alleviate it effectively. According to our current understanding, it’s agreed that noxious stimuli which are painful to humans also cause pain in other mammals. Although this doesn’t mean that they experience pain in the same way, it does imply that they do experience the aversive nature of pain the way we do.

Subjectivity prevents scientists from treating pain as a universal phenomenon. Animals born to mothers experiencing high stress levels while pregnant, for example, show amplified pain responses. Similarly, a mother’s neonatal experience of pain also affects her offspring’s response to pain. Interestingly, although pain responses in males and females are the same at birth, males tend to have a reduced sensitivity to pain as they age. The authors highlight that other factors, such as the animal’s personality, the presence or lack of social support, whether the animal has had previous experience with the specific type of pain, or even if there is a human present, can significantly affect how an animal deals with and responds to pain. This variability calls for individualistic and real-time methods for pain identification. 

Whichever technique will be adapted as the standard, it has to be valid, reliable and feasible. The researchers warn that real-life pain has many dimensions – the method should be able to consider the intensity, frequency, duration, and quality of the pain. It must also be useable on farms, in vet ORs, at home, in the field ,or within a laboratory animal facility, and yield identical results consistently. On top of that, all of this should ideally be measured in real-time and without the use of special equipment. Methods including spontaneous behaviour observation and measuring cortisol concentrations are the other most frequently used methods at the moment; while the former is not systematic enough to ensure reliability, the latter fails in validity, as the animals need to be handled in order to take the necessary blood samples, and often experience stress, which further alters their pain response.

On the other hand, a number of facial expression identification systems have already been developed for many animals, including rodents, rabbits, ferrets, cats, sheep, pigs, and horses, detailing all possible individual facial movements that can occur across the face. Although they have not yet been applied to pain determination, the prospect is promising, as both the sensory and emotional components of pain have been demonstrated in facial expression. One of the main problems this technique faces is that individual animal features such as hair length, muscularity, and coat color impair the observers’ ability to assess facial expressions. Furthermore, the scales need to be age-specific, as young animals’ reactions to pain differ from those of their adult counterparts. Finally, other emotions such as fear, or various illnesses, are also reported to skew pain expression.

The researchers advise that animal care staff need to be properly trained to use appropriate scales for the species under their care. Although the facial expression method might prove to be indispensible in the future, there is a need for the continued development of the current scales. Many of the scales need feasibility testing and reliability refinement before they can safely be used in clinics and in the field. In any case, animal advocates may start encouraging their local vets and governmental animal care institutions to look into these methods, the use of facial expression scales, which is both non-invasive and promising.