Anxiety or intuition: a more accurate question to ask

The common version of this question is that anxiety and intuition feel different and the work is to learn to tell them apart. Anxiety is the loud, urgent, fearful one. Intuition is the calm, clear, quiet one. With enough practice, the advice goes, you can listen to your body and know which is which.

The research does not really support this. The body does not have two settings. What it has is a single predictive system that runs in the background, generating expectations about what is about to happen and what is happening inside it, and producing the felt sense of being in a particular state. Anxiety and intuition are not two distinct signals coming from two distinct places. They are two different ways the same machinery can run.

The useful question is not which one you are having. It is what the machinery has to work with. The strength of a gut feeling does not tell you whether the underlying pattern is reliable. Whether the gut feeling carries information depends on how the patterns it is drawing on were learned, and on the state the system is in while you are trying to read it. Both of those are answerable questions. The clean separation by feel is not.

Interoception is the term for the brain's sense of what is happening inside the body. The everyday picture of how this works is that signals travel up from the heart, the gut, the muscles, and the brain reads them. The current research suggests the picture is closer to the reverse. The brain runs a continuous set of predictions about what the body is going to be doing, compares those predictions to the signals coming up, and the felt experience of being in your body is largely the prediction, adjusted by the gap between what was expected and what arrived.

Barrett and Simmons set this out in the Embodied Predictive Interoception Coding model, published in Nature Reviews Neuroscience. Greenwood and Garfinkel, in a 2025 review in the Annual Review of Psychology, traced how this model has held up across the field. The body is not just sending signals. The brain is asking questions of the body, then registering what does not match what it expected.

This matters for the anxiety-versus-intuition question because it locates both states in the same machinery. A gut feeling and a wave of anxiety are not produced by different organs reporting different things. They are produced by the same predictive system, running in different ways, against patterns that have been learned over time. The felt quality of the signal is the output. It is not the data.

Anxiety is what the predictive system looks like when its predictions about threat are weighted heavily and held with high precision. The brain is not making things up. It is generating expectations about what is dangerous, treating those expectations as carrying a lot of information, and producing bodily states that match. A racing heart, a tight chest, a sense of urgency: these are the system pre-loading the body for a threat it has predicted.

What this does to the signal is counter-intuitive, and it is worth being clear about. The popular idea is that anxious people feel their bodies more accurately, just with too much intensity. The meta-analytic evidence suggests otherwise. Adams and colleagues, in a 2022 systematic review and meta-analysis of fifty-five studies in Neuroscience and Biobehavioral Reviews, found no association between anxiety and cardiac interoceptive accuracy. People who score high on anxiety are not, on average, better at detecting their own heartbeat than people who score low.

What anxiety does change is something different. Clemente, Murphy and Murphy, in a 2024 meta-analysis of seventy-one studies, found that anxiety was associated with more frequent noticing of bodily signals, greater sensitivity to those signals, and a more negative evaluation of them. It was not associated with better use of bodily signals to inform emotional understanding. The system is louder. It is not more accurate.

Put simply, anxiety produces internal signals that feel important. The feeling of importance is generated by the predictive system itself, not by the underlying information. Confidence and accuracy are separate things. A gut feeling can be intense and certain because the machinery is running with high precision on a threat prediction, while the underlying pattern it is matched to has very little to do with the situation in front of you.

Intuition, in the research literature, is not a mysterious second sense. It is pattern recognition that has become fast enough to feel automatic. The system has encountered a situation many times, has registered which features matter and which outcomes followed, and now produces a quick judgement about a new instance without consciously walking through the comparison. That judgement arrives as a feeling rather than as an argument because the work happened below the level of deliberate thought.

This is why intuition can be remarkably accurate in some domains and remarkably wrong in others. The mechanism is the same. What changes is the quality of the underlying pattern. An experienced clinician walking onto a ward and feeling that something is off about a patient before they can name what is wrong is the output of a system that has been trained on many comparable situations with feedback about what came next. The same is true of an experienced parent reading a tone in their child's voice that is different from the usual ones. The feeling arrives quickly because the work was done long before the present moment.

The same mechanism, run on patterns learned from very few examples or with no reliable feedback, will still produce a confident feeling. It will just be a feeling matched to a pattern that does not, in fact, predict much about the present situation. The feeling itself does not know the difference. The work of telling them apart is not done by attending more carefully to how the feeling feels. It is done by looking at what the pattern was built from.

Kahneman and Klein, in a paper in American Psychologist, spent some time arguing about whether intuition could be trusted before realising they were not really disagreeing. They agreed on two conditions. The first is what they called a high-validity environment: the situation has to contain regularities that are stable and learnable. The cues you are picking up actually relate, in some predictable way, to the outcomes. The second is adequate practice with feedback: you have encountered enough instances of the situation, and received enough information about how things turned out, for the system to have learned the regularities.

When both conditions are met, intuition can be remarkably good. When either is missing, intuition still produces a feeling, but the feeling is not tracking the world. The same person can have trustworthy intuition in one domain and unreliable intuition in another, depending on the conditions they have been operating under. The subjective quality of the feeling does not change with the conditions. Only its accuracy does.

This shifts the diagnostic question. The useful question to ask of a gut feeling is not whether it feels like anxiety or intuition. It is whether the situation in front of you is the kind your nervous system has had a real chance to learn, with feedback that confirmed or contradicted its predictions. If yes, the feeling is worth taking seriously. If no, the feeling is worth noticing, but it is not, on its own, evidence about the situation.

What follows is one structure for working with a strong gut feeling when you are not sure what to do with it. It does not tell you whether the feeling is right. It puts you in a position to decide more accurately. The sequence is regulate, check the conditions, look at what the pattern is matched to.

Regulating first is not optional. While the threat system is loud, the precision of the prediction is high and the room for considering anything else is small. This is the state in which almost any signal will feel decisive. Slowing breathing, noticing contact with the ground, looking around the room: these moves do not change what is true. They change the weighting the system is giving to its current prediction, which is what makes the next two questions answerable at all.

Checking the conditions is the question Kahneman and Klein named: have I encountered this kind of situation many times, with feedback? If the answer is yes, the gut feeling is more likely to be tracking something real. If the answer is no, the strength of the feeling is not telling you about the world. It is telling you something about the precision the system is giving its current prediction. Both are important to know. They are different things to know.

Looking at what the pattern is matched to is the last move. Sometimes the match is current and specific: this person is doing the thing the previous person did. Sometimes the match is older and broader: any situation in which someone is annoyed reads as the older danger from a long time ago. The feeling can be the same in both cases. The two situations call for different responses.

Sometimes the three moves above will not give you a clear answer. The system is offering a strong feeling, the conditions are partially met, the pattern match is partial. This is common, and it does not mean the work has failed. It means the situation is genuinely uncertain. There is a different question to ask in that case. It is not whether the feeling is right. It is whether the feeling is asking for attention.

A signal that is persistent, that returns, that escalates over time, or that arrives with safety implications attached is worth attending to even if you cannot yet say what it is tracking. The current research on interoceptive interventions, including a 2023 systematic review of randomised trials by Heim and colleagues in Psychiatry and Clinical Neurosciences and Schuman-Olivier and colleagues' 2024 work on how trust in the body's signals shifts with structured practice, suggests that the relationship between a person and their internal signals can change with the right kind of attention, and that this shift sits underneath a lot of broader psychological work.

What that means in practice is that working out what to do with a recurrent gut feeling is rarely a one-conversation task, and is rarely best done alone. Bringing another mind into the question is not an admission that the feeling cannot be trusted. It is a way of giving the system more data than it can generate on its own.

The body does not have two settings. It has one predictive system, running with different precision against patterns of different quality. The strength of a gut feeling is not evidence about the world; it is evidence about the system. The work is not to learn which feelings to trust. It is to learn which conditions make a feeling worth trusting. That is a smaller and more answerable question.