Interoception

Why do people who cannot feel their own heartbeat also tend to misread their own emotions?

Sit still, close your eyes, and count your heartbeats without touching yourself. Some people can do this well; many cannot do it at all. And there is a claim attached to that difference which has become common knowledge: the poor counters are also the people who cannot tell irritation from hunger, or anxiety from excitement.

The claim is attractive because it has a mechanism ready-made. If an emotion is partly a reading of the body, then someone who reads the body poorly should read their emotions poorly. But notice how much that argument is asking the counting task to carry. Before accepting the conclusion, it is worth asking what the task actually measures.

Start with the sensible half. The body sends a continuous stream of information upward — from heart, gut, lungs, blood vessels — along vagal and spinal routes to brainstem and then cortical regions, the insula prominent among them. That much is anatomy, not theory. And there is good reason to think emotion is not simply accompanied by that stream but partly constituted by an inference over it: the brain has to work out what the state of the body means, and the same arousal can be read as fear, effort or attraction depending on what else is going on.

Now the sharp test. If emotional feeling really is an inference over bodily signals, the timing of those signals should matter, in a way no belief could produce. Cardiac signals are not continuous: baroreceptors fire in the brief window when the heart contracts and are quiet between beats. So present the same fearful face at systole and at diastole, and the two should be judged differently. They are — fear is reported as more intense at systole. That is a within-person manipulation the participant cannot detect. Had cardiac phase made no difference, or had it shifted judgements of emotionally neutral material just as much, the inferential account would be in serious trouble.

So the general idea survives. Return to the counting task and ask whether it does. Here the problem is severe. Someone who knows roughly what their resting pulse is can produce a plausible number and score well without detecting anything. Scores move when you change what people believe about their heart rate, and they correlate with body size. The task confounds knowing with feeling.

Which forces a distinction the field now makes routinely. There is accuracy — objective performance on a body-sensing task. There is sensibility — how much you say you attend to your body, by questionnaire. And there is awareness — whether your confidence in a judgement tracks whether that judgement was right. These three come apart, and the questionnaire measure, not the objective one, is the one that most reliably tracks anxiety.

Now the original claim can be examined properly. Alexithymia — persistent difficulty identifying and describing one's own feelings — is what the question is really about. Individual studies reported that it came with poor heartbeat detection, and the pairing entered general circulation as established. When the studies were pooled, the association with objective interoceptive accuracy came out close to nothing. The relationship that does hold, reasonably consistently, is with self-reported bodily attention — which is to say, with another questionnaire.

Three refinements, and the first is a correction of the question itself.

The premise is weaker than it sounds. On present evidence the honest statement is not "people who cannot feel their heartbeat misread their emotions" but "people who report attending poorly to their bodies also report difficulty naming their feelings, while the objective measure shows little relationship to either." Those are very different claims, and the second is close to a correlation between two questionnaires. The famous version is a case of a plausible mechanism being credited with evidence that belongs to a much narrower finding.

The second refinement rescues the mechanism from that verdict. The failure is of a measure, not of the idea. Bodily signals demonstrably shape emotional experience: cardiac timing modulates fear, drugs that blunt peripheral arousal reduce the intensity of reported fear, and panic can be provoked by manipulating carbon dioxide — a bodily route with no cognitive content at all. Interoception matters to emotion. What is not established is that a heartbeat-counting score tells you how well any individual does it.

The third is about direction. Even a clean correlation would not settle the causal arrow. Poor signal detection could impair emotional labelling; equally, a habit of avoiding one's bodily state — which is characteristic of both alexithymia and anxiety — could degrade detection, and a third factor such as attention could produce both. The interventional evidence, chiefly training studies claiming improved detection with emotional benefit, is small-sample and mixed.

• Why the cardiac cycle modulates fear specifically, and whether other emotions follow the same timing.
• How predictive accounts explain persistent bodily symptoms with no detectable peripheral cause.

Every term the collection defines is gathered in the glossary.