Biologists studied over 7,000 recordings from 411 pigs, from the brief squeaks of satisfaction at feeding time to the desperate cries at slaughter, before classifying them into 19 different categories.
“We show it’s possible, basically, to figure out the emotions of the pigs according to their vocalisations,” project leader Elodie Briefer, a lecturer at the University of Copenhagen, told AFP.
The project, split between Switzerland, Denmark, Norway, France and the Czech Republic and published in the journal “Nature”, offers a new way of improving animal welfare by laying the groundwork for a tool that can categorise an emotion based on the noise produced, according to the researcher.
“We also run a machine-learning algorithm that produces a spectrogram, which is then trained to recognise negative and positive contexts.”
Once developed, the new tool would allow farmers, who today can mostly only check the physical wellbeing of the animals, to monitor their mental health.
If the negative squeals increase, the farmer would be alerted that something was wrong, Briefer said.
The Scandinavian country is home to 13.2 million pigs, making it the leader in Europe with over two per capita. So, for the Danish Agriculture and Food Council, the implications of the study are promising.
“This concept could potentially be a useful tool among others in the work to monitor the health and wellbeing of pigs,” a spokesperson said.
‘Very vocal’
According to Briefer, they reached “92% accuracy of classifying the valence – whether the call is negative or positive – and 82% accuracy in classifying the actual context in which the sounds were produced”.
Positive feelings are expressed in short grunts, while negative sentiments are most often expressed with longer sounds, the findings revealed.
But why focus on pigs rather than cows or rabbits? For the authors, the pig, known for its wide range of squeaks and noises, was the perfect match.
“They’re very vocal, which makes them easier to study,” Briefer said. “They produce vocalisations all the time, and even in a low-intensity situation, they would still vocalise.”