Strategy-dependent causal relationships between large-scale cortical activity and perceptual decisions

Every day, an animal faces a simple but critical question: which way should I go? To answer this, it must gather clues from the world around it, weigh them against what it has learned, and make a choice. For decades, scientists studying these decisions have often assumed that the brain operates like a single, steady machine, processing information in the same way every time a choice is made. But in reality, the mind is far more fluid. Just as a person might sometimes focus intently on a puzzle and other times rely on a hunch or a habit, animals switch between different internal strategies to solve the same problem. Sometimes they follow the evidence carefully; other times, they fall into a pattern of favoring one side over the other, regardless of the clues. The big question has been: does the brain change its wiring to support these different modes of thinking, and if so, which parts of the brain are responsible for keeping the animal on the right track?

A team of researchers set out to map this hidden landscape of decision-making in the mouse brain. They turned their attention to a specific task where mice navigate a virtual T-shaped maze. As the mice run forward, they see a series of visual towers appearing on either the left or right side of the path. The goal is simple: remember which side had more towers and turn toward it to get a reward. The researchers gathered data from dozens of mice, watching them perform this task over thousands of trials. To understand how the brain works, they didn't just watch the mice; they also used a technique called optogenetics. This method allows scientists to temporarily silence specific groups of brain cells with a flash of light, effectively turning off a switch in the brain to see what happens when that part stops working. By combining these light-based interruptions with recordings of brain activity, the team could test how different brain regions contribute to the decision-making process.

The first step was to figure out what kind of thinking the mice were doing at any given moment. The researchers analyzed the choices the mice made when they were not being interrupted by the light. They discovered that the mice were not using just one fixed strategy. Instead, they drifted between three distinct mental states. In the first state, the mice were fully engaged, paying close attention to the towers and letting the evidence guide their choice. In the other two states, the mice were disengaged. In these moments, they largely ignored the towers and instead developed a strong habit of turning either left or right, regardless of what they saw. These states were not fleeting; once a mouse entered one, it tended to stay there for many turns in a row, sometimes for more than thirty trials. This finding confirmed that the brain does not just process data; it adopts different modes of operation that persist over time.

With these three states identified, the researchers then asked what happened when they silenced different parts of the brain. They targeted dozens of locations across the top surface of the brain, known as the cortex, and watched how the mice's choices changed. The results were striking and depended entirely on which mental state the mouse was in. When a mouse was in the engaged, evidence-driven state, silencing the front part of the brain caused it to lose its focus. The mouse stopped listening to the towers and began to guess, often developing a strong bias toward one side. However, when the mouse was already in a disengaged, biased state, silencing the same front part of the brain had the opposite effect. It actually reduced the mouse's pre-existing tendency to favor one side, making the bias less extreme. This paradoxical result showed that the front part of the brain acts as a control center that is essential for maintaining focus when the animal is trying to do the right thing, but its removal can sometimes reduce the strength of a distraction, even if it does not fully restore the animal's ability to use evidence.

The researchers also looked at the timing of these interruptions. They found that silencing the front of the brain had the strongest impact when it happened later in the trial, just before the mouse had to make its turn. This suggests that the front brain regions are crucial for holding the decision in mind and executing the final choice, especially when the animal is trying to be accurate. In contrast, when the researchers silenced the back part of the brain, the effects were different. These areas seemed to play a larger role when the animal was in a biased, disengaged state, perhaps helping to maintain those habits or process the history of past rewards.

To see the whole picture, the team also recorded the electrical activity of the brain across its entire surface using a widefield camera. They found that the front part of the brain was much more active when the mice were in the engaged state compared to when they were biased. Furthermore, when they tried to predict what the mouse would choose based on brain activity, the front regions were the best predictors during the engaged state, while the back regions were more informative during the biased state. This confirmed that the brain reorganizes itself depending on the strategy being used. The front of the brain is the key to staying focused and making a choice based on the current evidence, while other areas take the lead when the animal falls back on habits.

These findings suggest that the brain is not a static machine with fixed parts doing fixed jobs. Instead, it is a dynamic system that reconfigures its connections and activity patterns depending on how the animal decides to approach a problem. The front part of the brain emerges as a critical hub for maintaining an engaged, evidence-based strategy, acting as a guard against the tendency to fall into simple biases. When this guard is down, the animal loses its way, but interestingly, removing it can sometimes reduce the intensity of a distraction, though not necessarily return the animal to optimal performance. This work provides a clearer view of how large-scale brain dynamics shift to support different ways of thinking, revealing that the path to a decision is as much about the brain's current strategy as it is about the evidence in front of it.