When we try to understand an animal’s behavior, it’s tempting to think in simple terms. If something good happens, they feel happy. If something bad happens, they feel fear or pain. But the truth is more complex. This article explains an important theory that arose from studies of human cognition and neuroscience — the “Bayesian brain hypothesis” — and argues that it applies to animals too. If the authors are correct, it could change how we understand animal behavior, welfare, and suffering.

The Bayesian brain hypothesis, sometimes called the “predictive brain” model, suggests that individuals don’t react to events naively or automatically. Instead, they constantly compare what’s happening around them to what they expect based on past experiences. In this way, the brain works like a prediction engine. It builds internal models of the world, and then uses those models to anticipate what will happen next. New information, coming in through the senses, is used to update these models under normal circumstances.

These processes happen mostly outside of conscious awareness. The placebo effect in humans provides a good example. When people are led to believe they’re taking a medicine that will reduce their pain, they expect less pain — and that’s often their experience. The Bayesian brain hypothesis argues that this predictive process is at the core of how human brains work. But what about animal brains?

There’s growing evidence that animals operate in similar ways. For instance, one study found evidence for a placebo effect in rats. The researchers first injected the rats with a pain-reducing medicine, and this led to behaviors consistent with the experience of less pain. Later, they injected them with a non-active compound, and the rats showed behavior similar to their responses to the first injection. Apparently, the rats had learned to anticipate pain reduction from an injection. Based on their past experience, the rats behaved as if they expected pain relief.

The Bayesian brain hypothesis also offers insight into why animals explore their environments. Exploration, even when risky or tiring, helps animals update their internal models of the world. Doing so reduces uncertainty, which can help them better predict what’s likely to happen in the future. Often, animals may even revisit familiar areas not out of habit, but to test and refine their expectations — a process the Bayesian brain hypothesis helps explain.

The authors argue that the Bayesian brain hypothesis can help us understand how animal suffering and poor welfare can arise. If an animal’s past experiences are consistently negative — if they endure pain, isolation, or fear — they may develop pessimistic expectations about the world. As a result, they may come to “expect” continued suffering, and may even stop trying to escape from bad situations. This is similar to the concept of learned helplessness in humans, where individuals behave as if they’re powerless to change their situation. Under such situations, the typical model-updating process may be disrupted, leaving one stuck with very negative, seemingly unchangeable expectations.

This may explain why animals who are kept in poor conditions for long periods often show signs of emotional distress: reduced interest in rewards, heightened fearfulness, and depression-like symptoms. These signs suggest not only that the animal is suffering in the moment, but that their internal models of the world have shifted toward hopelessness.

The authors don’t claim that the Bayesian brain hypothesis explains everything about animal behavior. But they argue that it offers a helpful framework for thinking about how animals learn and adapt to their environments, and how this may impact their current feelings. This has important implications for animal advocacy and welfare science.

For advocates, the Bayesian brain hypothesis can offer a powerful tool to explain why animals in captivity may act in seemingly “passive” or “strange” ways: not because they don’t care, but because they’ve learned not to expect anything better. Viewing animals as predictive, learning beings highlights the importance of consistently providing them with safe, enriched environments and careful handling practices. For researchers, the authors encourage more focus in future studies on how positive experiences — like enrichment or gentle handling — can help animals develop more hopeful expectations.

Suffering can arise not only from what one experiences in the moment, but also from what one anticipates. Reducing harm means responding to both in the animals we care for.