Frontal population organization supports the behavioral expression of confidence
This study demonstrates that in rats, the behavioral expression of robust confidence representations in the orbitofrontal cortex depends not on the strength of individual neuron activity, but on the specific temporal and geometric organization of the neuronal population, which determines whether these internal states can be accessed and translated into action.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
Inside the brain, thoughts often exist as quiet, internal signals that do not always translate into action. A person might feel absolutely certain about a choice, yet hesitate before acting, or feel unsure and rush ahead anyway. For decades, scientists have known that specific brain cells can encode feelings of confidence, firing more or less depending on how sure an animal is about a decision. However, a fundamental mystery remained: just because a brain cell carries a signal of confidence, does that guarantee the animal will act on it? The question is not merely about whether the brain knows the answer, but how that knowledge becomes a physical behavior. This research explores the gap between having a thought and expressing it, investigating why a strong internal feeling sometimes fails to guide a rat's actions.
To solve this puzzle, researchers at RIKEN and Doshisha University designed a task for rats that separated the moment of decision from the moment of reward. The animals were trained to sniff a mixture of two scents and choose a port on the left or right to find water. The difficulty of the scent mixture varied, making some choices easy and others hard. After making a choice, the rats had to wait for a reward that might or might not arrive. In one version of the experiment, the reward arrived at a completely unpredictable time, forcing the rats to rely on their own judgment of how likely they were to be correct. If they felt confident, they waited longer; if they felt unsure, they left sooner. In this scenario, the time the rat chose to wait was a direct measure of its confidence.
The researchers then introduced a second version of the task where the reward timing was fixed and predictable. The water always arrived exactly half a second after the choice. In this condition, the rats' perceptual decisions remained sharp; they still chose the correct port with the same accuracy as before. However, their waiting behavior changed dramatically. They no longer waited longer when they were confident. Instead, they waited for a similar amount of time regardless of whether they felt sure or unsure. The internal feeling of confidence was still there, but it had lost its power to shape the animal's behavior. This created a perfect opportunity to ask: did the brain stop feeling confident, or did it simply stop using that feeling to guide the wait?
To find out, the team recorded the activity of hundreds of neurons in the orbitofrontal cortex, a region of the brain known for processing value and confidence. They discovered that the neurons themselves had not changed. Even when the rats stopped using their confidence to decide how long to wait, the individual brain cells continued to fire in patterns that clearly signaled high or low confidence. The internal representation of the feeling was robust and intact. The problem was not that the brain had forgotten how to feel confident; the problem was that the brain had stopped organizing those feelings in a way that could be easily read by the rest of the body.
The researchers found that the key difference lay in how the population of neurons worked together. When the rats were in the unpredictable condition and their confidence guided their waiting, the neurons fired in a highly organized, stable pattern over time. This organization allowed the brain to maintain a clear, linear signal of confidence that could be accessed at any moment during the wait. It was as if the neurons were speaking a consistent language that the rest of the brain could understand and act upon. In the predictable condition, where confidence no longer guided behavior, this organization fell apart. The neurons still fired, but their collective pattern became disordered and less stable. The signal was no longer arranged in a way that allowed for a smooth, linear readout. The information was present, but it was buried in a chaotic structure that the downstream parts of the brain could not easily access.
To prove that this organization was the missing link, the scientists used light to temporarily disrupt specific connections between the orbitofrontal cortex and the striatum, a deeper brain structure involved in movement and habit. When they inhibited or activated these specific pathways during the waiting period, they could break the link between confidence and waiting time. Even in the unpredictable condition where the rats usually waited based on confidence, disrupting these connections caused the rats to stop adjusting their wait times according to their certainty. This confirmed that the orbitofrontal cortex does not just hold the feeling of confidence; it must also organize that feeling into a specific pattern to send it downstream.
The study reveals that having a strong internal representation is not enough to drive behavior. The brain must also arrange its signals in a specific, organized geometry that makes them accessible to the rest of the system. When the brain's population of neurons is well-organized, internal thoughts like confidence can flow freely into action. When that organization is disrupted, even a clear and strong thought remains trapped inside, unable to influence what the animal does next. This suggests that the flexibility of behavior depends less on how strongly we feel something and more on how the brain arranges those feelings to be read and acted upon.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.