Researchers utilized machine learning to observe how mice manage aggressive interactions with one another. Their results revealed that male mice can reduce such aggression by approaching a female mouse to divert the attention of the aggressive male.
A team of researchers led by Joshua Neunuebel at the University of Delaware, USA, utilized machine learning to study mice behavior in response to aggression from other mice. Their findings, published on October 15th in the open-access journal PLOS Biology, indicate that male mice mitigate aggressive confrontations by running toward a female mouse to distract the aggressive male.
The researchers observed interactions among groups of two male and two female mice over a span of five hours. Mice, like many other species, establish social hierarchies, and in nearly every group they recorded, one male was notably more aggressive toward the other.
Studying social interactions can be quite complex, so the researchers employed a machine learning technique to analyze the aggressive behaviors and the subsequent responses of the mice. They tracked more than 3,000 aggressive incidents between the male mice, and the machine learning system helped identify the most common responses to aggression and whether those responses either resolved or escalated the situation.
The study revealed that the male mouse who was subjected to aggression often darted toward one of the female mice, which successfully deescalated the conflict. This behavior could be seen as a “bait-and-switch” strategy, as the initially aggressive male typically pursued the other male but ended up interacting with the female instead of continuing the aggressive encounter.
While some other strategies might avoid conflict temporarily, they could later lead to full-blown fights. However, after employing the bait-and-switch approach, the researchers observed that fights were rare; the male mice tended to keep their distance, with the aggressive one focusing his attention on the female mouse.
Although this bait-and-switch tactic effectively reduces conflict, it may pose challenges for the victim, such as losing time spent with female mice. Future research may explore the effectiveness of these strategies in larger groups. This study illustrates how machine learning tools can enhance our understanding of animal behavior, and similar methodologies could be applied to investigate how various species with social hierarchies address aggression.
The authors noted, “Through the use of artificial intelligence, we discovered that male mice seek out nearby females to distract aggressors and diffuse conflicts. Following an aggressive incident, the mouse that was attacked briefly interacts with the female before quickly escaping, as the aggressor’s attention shifts toward her.”
