Study shows how machine learning can identify social grooming behavior from acceleration signals in wild baboons

Researchers from Swansea University and the University of Cape Town have monitored social grooming conduct in untamed baboons by employing accelerometers mounted on collars.

The research, which was published in the Royal Society Open Science journal, is the initial study to effectively measure grooming budgets utilizing this technique, thereby unveiling a plethora of future research prospects.

By utilizing collars that contained accelerometers produced at Swansea University, the group documented the undertakings of baboons in Cape Town, South Africa. They identified and evaluated general activities such as strolling, foraging, resting, and running, as well as the exchange of grooming between individuals.

An algorithm for supervised machine learning was instructed on acceleration data that was synchronized with video recordings of baboons, and it was able to accurately distinguish the exchange of grooming between baboons with a high level of accuracy.

The group subsequently employed their machine learning model to acceleration data gathered from 12 baboons to measure grooming and other activities constantly, day and night.

Dr Charlotte Christensen, the lead author from the University of Zurich, stated, “We were uncertain whether a collar-mounted sensor could identify a behavior that includes such subtle movements, but it was successful. Our discoveries have significant implications for examining social behavior in animals, particularly non-human primates.”

Social grooming is among the most significant social behaviors in primates, and it has been a key area of focus in primatology research since the 1950s.

Previously, scientists had depended on observing primates directly to determine the extent to which they groom each other. While direct observations offer systematic data, it is sparse and non-continuous, with the added restriction of researchers only being able to monitor a few animals at a time.

The technology employed in this study is revolutionizing the field of animal behavior research and enabling innovative new areas of exploration.

Dr Ines Fürtbauer, the senior author from Swansea University, stated, “Our team has wanted to accomplish this for years. The capability to gather and analyze continuous grooming data in wild populations will enable researchers to revisit longstanding questions and explore new ones regarding the establishment and maintenance of social bonds, as well as the mechanisms that underlie the sociality-health-fitness relationship.”

Read the paper in full: Quantifying allo-grooming in wild chacma baboons (Papio ursinus) using tri-axial acceleration data and machine learning

-END-