A pig’s grunting and squeaking reflects whether it feels “good as hell” – or is stressed, bored or scared: With the help of artificial intelligence, researchers have succeeded in decoding the emotional meaning of pig noises. The scientists say that the findings could now be used to improve animal welfare in pig farming.
How do pigs and co feel with the sometimes depressing housing conditions in animal husbandry? Basically, it is now accepted that the emotional state is also an important part of animal welfare. But the focus is still on the physical condition of livestock. In order to automatically monitor the physical condition of the animals, farmers also have several analytical systems at their disposal. In contrast, one might think that it is hardly possible to obtain objective insights into the animal world of sensations. But as the current study shows in the case of pigs, conclusions can be drawn – by listening to the animals carefully.
Basically, it should be emphasized that pigs are neither stupid nor undemanding: They are highly developed mammals with complex cognitive abilities. This is also reflected in their complex vocalizations. So far, however, little is known about their importance. Therefore, an international team of scientists has devoted itself to the systematic research of the emotional “pig language”. To do this, they used artificial intelligence methods: They trained an adaptive system to recognize connections between certain sound characteristics and emotional states in pigs.
Can you hear the emotional state?
The basis for training the system was formed by more than 7000 sound recordings collected in the commercial and experimental environment. They emerged in a wide range of positive and negative situations that animals may face, from birth to death. Positive situations included, for example, when piglets are being suckled by their mothers or when they are reunited with their families after isolation. Emotionally negative experiences included breakups, food deprivation, and conflicts, among others. The sounds made during the extremely negative experiences were also part of the study: during castrations and when handling the animals in the slaughterhouse.
In experimental pens, the researchers also presented the pigs with different environments designed to evoke more differentiated emotions in the middle of the spectrum. This included an arena with toys or food and a comparison environment without any stimuli. The researchers also placed new and unfamiliar objects in the arena for the pigs to interact with. Meanwhile, the pigs’ calls, behavior and heart rate were monitored and recorded. The researchers then fed the more than 7,000 sound recordings into their ‘smart’ computer system to see how well they could identify a pattern in the sounds depending on emotional conditions.
On the way to a monitor system
As the researchers report, their approach turned out to be successful: They managed to develop an algorithm that can use the sounds to decode how a pig is emotionally feeling. “There are clear differences in what the pigs say when we look at positive and negative situations,” says lead author Elodie Briefer from the University of Copenhagen. “In positive situations, the sounds are much shorter, with little variation in amplitude. Grunts start out high and gradually decrease in volume,” says Briefer. Based on these sometimes subtle features, the trained algorithm can now correctly assign a large part of the sounds, the researchers report. They also match other signs: with negative emotions, pigs tend to stand still or try to withdraw, while positive emotions include exploratory behavior and pricking up their ears.
The scientists now see a basis for further developments in their study results: “We prove that noises made by animals can give a good insight into their emotions. This is an important step towards improved animal welfare in livestock farming,” says Briefer. Specifically, she and her colleagues hope the approach could evolve into a system for automatically monitoring pigs, allowing farmers to keep an eye on their animals’ mental well-being. However, this requires further development work: “So far we have trained the algorithm to decode the grunt. Now we need someone to integrate it into an app that users can use to improve the welfare of their pigs,” says Briefer.
Source: University of Copenhagen, professional article: Scientific Reports, doi: 10.1038/s41598-022-07174-8