Animal welfare: a translator to decode the emotions of pigs through their growls


You may also like this partner content (after the ad)

Today, the notion of animal welfare is everywhere: in the expectations of consumers and citizens, in specifications, in regulations, in marketing initiatives, in farms, etc. Animals are recognized today as sentient beings. It is a scientific achievement, but also a legislative reference that responds to a strong demand from society. And, just like humans, animals use sounds to convey their emotions, negative or positive. In this context, the objective of the research is practical: to understand the mental health of animals kept in captivity, in order to improve their well-being and their living conditions. Recently, an international team of scientists has developed an algorithm that can decode the grunts of pigs, from birth to slaughter. This discovery marks a significant breakthrough for animal welfare. This decoder is a valuable and reliable non-invasive tool for assessing pig emotions and monitoring their welfare in real time.

In his book “ The expression of emotions in humans and animals ”, Darwin describes many species using sounds to convey their emotions. He insists on the fact that animals that are usually not very talkative, such as rabbits, produce shrill cries when they are in pain. He explains that many also vocalize in situations of well-being, when they find a lost companion for example.

Animal emotions, defined as Intense, short-term reactions to specific events have attracted increasing interest in recent decades, not least because of growing concern for animal welfare. But what are the acoustic markers of these emotions, how are they coded? In order to answer these questions, a team of international researchers focused on domestic pigs. Their results are published in the journal Scientific Reports.

Recognizing the presence of emotions in animals: starting point

Just like humans, animals experience fear of novelty and frustration when the situation does not meet their expectations. They are reassured by predictable events and enjoy being in control of their environment. The knowledge acquired on the mother-young relationship and, more generally, on social ties, as well as the demonstration of their cognitive abilities, also contribute to bringing animals closer to humans. Neurobiology, for its part, shows that in mammals, similar cerebral structures are involved in different emotional behaviors.

Alain Boissy, director of research at INRA-Clermont-Ferrand-Theix and co-author of the analysis, declares: “Even if it is still difficult to have operational indicators of their mental state, we have made considerable progress in the objectification of the emotions of animals ”. These advances have been made concrete by the European program Welfare Quality, which has established in 2000 a European standard for the assessment of animal welfare. It includes measurements on the animal, its health and its behavior, in addition to measurements on the breeding conditions.

It should be noted that the research at the animal confirms that emotions are not automatic and reflexive processes, but rather can be explained by elementary cognitive processes. This line of thought suggests that an emotion is triggered by an individual’s assessment of their environmental situation. This is the case of domestic pigs displaying very sophisticated varieties of vocal expression.

In fact, previous studies have established correlations between high-frequency calls, such as screams, associated with negative emotions, and low-frequency growls associated with positive or neutral emotions. Elodie Briefer, professor at the University of Copenhagen, and co-author of the study specifies, in a press release: “An emotion can be described by two main dimensions: its valence (positive or negative) and its level of intensity ”. But between these two extremes lies an assortment of less well-understood sounds. Indeed, coding the emotions in his vocalizations is one thing. Being able to “decode” them when we hear acoustic signals is another.

Varied situations, contrasting emotions, modulated sounds

For this purpose, the researchers recorded pig sounds in various scenarios, associated with positive and negative emotion. Positive situations included for example those where the piglets are suckling their mothers or when they are reunited after being separated. Emotionally negative situations included, among others, separation, fights between piglets, castration and slaughter. Through experimental parks, the researchers also designed in-between situations to evoke more nuanced emotions. These included an arena with toys or food and a similar arena without any stimulus. In addition, pig calls, behavior and heart rates were monitored and recorded where possible.

experience nuance emotion vocalization pigs
Photographie des zones expérimentales où des scénarios plus nuancés étaient testés. ©E. F. Briefer et al., 2022

” data-image-title=”experience-nuance-emotion-vocalisation-porcs” data-large-file=”” data-lazy-sizes=”(max-width: 633px) 100vw, 633px” data-lazy-src=”” data-lazy-srcset=” 633w,×300.jpg 284w,×528.jpg 500w,×499.jpg 472w,×412.jpg 390w” data-medium-file=”×300.jpg” data-orig-file=”” data-orig-size=”633,669″ data-permalink=”” height=”669″ src=”data:image/svg+xml,%3Csvg%20xmlns=’’%20viewBox=’0%200%20633%20669’%3E%3C/svg%3E” width=”633″>

Photography of experimental areas where more nuanced scenarios were tested. © EF Briefer et al., 200020669

Subsequently, the researchers analyzed the 7414 recordings of the 411 pigs, in order to establish a pattern in sounds based on emotions. In other words, the algorithm would be the key to decoding the pig’s emotions in order to know whether it experiences a positive (“happy” or “excited”), negative (“scared” or “stressed”) or intermediate emotion. .

In general, the results validated previous observations. The team observed more high-frequency calls (such as shouting and squealing) in negative situations. At the same time, low-frequency calls (such as barking and growling) occurred in situations where the pigs experienced more positive emotions. Apparently, two particular acoustic characteristics have proven to be as important as frequency in understanding emotional valence: duration and rate of amplitude modulation.

Indeed, previous studies conducted by Elodie F. Briefer and her colleagues have highlighted a relationship between the physical structures of sounds and the motivation underlying their use (joi e of breastfeeding, anxiety about a new situation, etc.). This principle was first stated in 500, by Eugène Morton, researcher emeritus at the Smithsonian Institution in Washington: the acoustic structure of a sound signal reflects the motivation of the sender (“motivation-structural rules ”). In other words, the vocalizations are louder and produced at a faster rate, higher pitched and more modulated in frequency, when the intensity of the emotion increases. EF Briefer clarifies: “There are clear differences in hog calls when we look at positive and negative situations. In positive situations, calls are much shorter, with minor fluctuations in amplitude. Growls, more specifically, start high and gradually decrease in frequency . Finally, these sound variations correspond closely to the effects of anatomical and physiological changes (complex vibrations of the vocal cords, involuntary contraction of muscles, etc.), precisely linked to variations in intensity. This is how, when an individual experiences a strong emotion, his voice becomes more acute.

A decoding key for farmers

The researchers then used a neural network to develop an algorithm able to translate the emotional characteristic of pig sounds. EF Briefer explains: “By training an algorithm to recognize these sounds, we can associate 92% of calls to the right emotion”. Additionally, the researchers were able to classify the pigs’ emotions according to how they naturally react to various stimuli. For one thing, typical signs of negative emotions in pigs include immobility, lots of vocalizations, and attempts to escape. On the other hand, signs of positive emotions include exploring their surroundings and forward ear position.

Today , it is widely accepted that the mental health of cattle is important to their overall well-being. Nevertheless, animal welfare today focuses primarily on the physical health of livestock. Indeed, several systems exist allowing to monitor automatically, only the physical health of an animal, for a breeder.

This is why the researchers hope that their algorithm could pave the way for a new platform allowing farmers to monitor the psychological well-being of their animals. EF Briefer says: “We trained the algorithm to decode pig grunts. Now we need someone who wants to develop the algorithm into an app that farmers can use to improve the welfare of their animals “. Especially since the suffering of the animals reflects on the breeders. Establishing a positive relationship with the animal, through specific practices, based on a better understanding of animals, gives meaning to the profession. It is an innovative concept called “One welfare“: a well- being interdependent between man and animal.

Animal welfare, between utopia and reality

Obviously, improving farm animal welfare can result in increased production costs (increasing building areas per animal, for example). Consumers in developed countries, particularly in Europe, say they are ready to bear this additional cost at the time of purchase. However, this desire for animal welfare must compete with cheaper products from countries that are less demanding in terms of animal welfare.

According to INRAE ​​researchers, animal welfare is a common good that is insufficiently taken into account by the markets. This state of affairs justifies intervention by public authorities at the international level, or failing that, the application of border adjustment mechanisms, including within the European Union, in the event of differences in regulatory standards between States. members.

The authors conclude that with enough data to train the algorithm, the method could also be used to better understand the emotions of others animals such as cattle or sheep. But are the states ready to assume, economically, such a scientific advance?

Source: Scientific Reports