House mice (Mus musculus domesticus) demonstrate prosocial behaviour only in the absence of food rewards

In the animal kingdom, kindness is not a universal trait. While some species are known to help one another without expecting a return, others act primarily in their own self-interest. Scientists have long wondered what drives this difference. One leading idea suggests that animals which raise their young together, sharing the burden of feeding and protecting offspring, might naturally develop a stronger tendency to help others. This concept, known as the cooperative breeding hypothesis, proposes that the daily practice of caring for another's young trains an animal to be more generous in other situations as well. To test this, researchers often look at how animals behave when faced with a choice: help a companion or help themselves. However, most of what we know comes from studying animals in laboratories, which may have lost some of their natural instincts through generations of breeding. To truly understand the roots of kindness, scientists need to observe animals that live more like their wild ancestors, where survival depends on making real choices in a complex world.

A team of researchers set out to investigate this question using the house mouse, a small rodent that lives in groups and shares nesting duties. They wanted to see if these mice would act kindly toward a companion when food was involved, and if their behavior changed when food was not on the line. The team worked with twenty female mice that were descendants of wild populations caught in Iran and Kazakhstan, keeping them in a laboratory setting that mimicked their natural environment as closely as possible. These mice were not the standard, highly inbred strains often used in labs, but rather animals that retained the genetic diversity and behavioral tendencies of their wild relatives. The researchers trained these mice to participate in two different experiments designed to measure their willingness to help.

The first experiment was a test of choice involving food. The mice were placed in separate but connected compartments where they could see and smell each other but could not touch. One mouse, the "actor," could pull a lever to slide a tray of food toward itself. Depending on how the tray was set up, pulling it could bring food only to the actor, only to the partner, or to both of them. The researchers wanted to see if the mouse would choose the option that helped its partner, even if it meant getting less food for itself. Out of the twenty mice trained, only six learned the task well enough to be tested. The results were surprising. When the mice had to choose between getting food for themselves or for their partner, they almost always chose themselves. Even when they could get food for both of them without losing anything for themselves, they did not show a strong preference for the shared reward over taking it all for themselves. In fact, when the choice was between helping a partner or getting nothing, the mice were no more likely to help the partner than they were to drop food into an empty cage where no one was present. It seemed that in the context of food, these mice were focused entirely on their own needs.

The researchers then tested the same six mice in a completely different situation, one that had nothing to do with eating. This was a release test. The partner mouse was placed inside a small, clear trap with a door that could be opened from the outside by pulling a bar. The actor mouse was free to move around the room. The researchers watched to see how quickly the actor would open the trap to free the trapped mouse. They compared this to situations where the trap was empty, held a novel object, or contained a piece of fresh banana, a high-value treat. The results here were the opposite of the food test. When the partner was trapped, the mice opened the door significantly faster than they did in any of the other situations. They were quicker to free a friend than to investigate a new object or even to get to a piece of delicious banana. Once the door was open, however, the mice did not immediately cuddle or groom their released friend. In many cases, the mouse that opened the trap simply walked inside the cage to explore it, leaving the freed partner behind. This suggested that the act of opening the door was not driven by a desire for social contact, but rather by a genuine concern for the trapped mouse's situation.

The study highlights that kindness in house mice is not a simple, all-or-nothing trait. Instead, it is highly dependent on the situation. When food is involved, these mice prioritize their own survival and do not seem to consider the needs of their companions. Yet, when a friend is in trouble and no food is at stake, they act quickly to help, even without a direct reward for themselves. This behavior challenges the idea that communal breeding automatically leads to high levels of generosity in all contexts. It suggests that while these mice are capable of empathy and helping, their instincts are finely tuned to the specific demands of the moment. The researchers found that the type of test used matters greatly; a test involving food might make an animal look selfish, while a test involving a trapped friend might reveal a deep capacity for care. By using wild-derived mice and testing them in different ways, the study provides a clearer, more nuanced picture of how prosocial behavior evolves, showing that even in a species known for living together, the drive to help is not constant but shifts with the circumstances.