Interoceptive exposure

Interoceptive exposure is the part of panic treatment that sounds like a mistake: the deliberate production of the bodily sensations a person is most afraid of. Spinning in a chair until the room moves. Breathing through a thin straw. Running up stairs. Over-breathing on purpose. Holding the breath, tensing everything, standing up fast from a crouch.

The usual description — getting used to the sensations — is close and slightly wrong, and the correction is the whole point. The target is not the sensation. It is the sensation's job as a forecast. Right now a racing heart arrives and announces that something terrible is beginning, and the exercise exists to produce it often enough, in enough different conditions, that it stops announcing anything. What changes is not how pleasant dizziness is. It is whether dizziness is news.

After a few attacks a second kind of avoidance forms, much quieter than the kind that shows on a map.

Exercise stops, or stays below the point where the chest is noticeable. Coffee goes. Then saunas, spicy food, dancing, sex, a sprint for a closing door. Each item is dropped for a reason that sounds sensible alone, and nobody adds them up.

The arrangement guarantees that the prediction never meets evidence. A belief that a heart at a hundred and forty means something is going badly wrong cannot be disconfirmed by a life organised so the heart never reaches a hundred and forty. And when the resting body is the only baseline available, any departure from it reads as a departure from safety: the normal range narrows while the fear stays the size it was.

Why relief teaches the wrong lesson is set out in safety behaviours and not repeated here — anxiety drops now, the learning that would have ended it is blocked, the credit goes to the precaution. The channel these sensations arrive through is interoception, used here only as a premise: a loud instrument under alarm and a badly calibrated one. The attack itself, its curve and its thirteen symptoms, is panic attack.

What this page adds is the asymmetry. Avoidance of the body hides better than avoidance of places. Someone who has stopped getting on trains looks exactly like that. Someone who has stopped running has an explanation available at all times, and the explanation is health.

The idea came from one observation: in panic the feared object is internal. Clark's 1986 cognitive model put the misreading of bodily sensations at the centre — a sensation is taken as evidence of catastrophe, which raises arousal, which produces more sensation. If that is the loop, what to expose someone to is not a place. It is their own chest.

Barlow and colleagues built the first structured treatment on that logic in 1989, with symptom induction at its core, and it beat relaxation alone and a waiting list. Its descendants are what the Cochrane network meta-analysis and the guidelines put first. But the component has always travelled inside the package rather than as a treatment of its own, and almost no trial isolates it — roughly where Boettcher, Brake and Barlow leave it: widely used, well grounded, under-studied alone.

Three findings make it look load-bearing.

It has been tested on its own. Lee and colleagues gave patients with panic disorder a brief intervention built around it and found reduced fear of the induced sensations alongside clinical improvement, specific to the sensations targeted.

Components are separable, and not all survive. Schmidt and colleagues dismantled the standard package and found breathing retraining added nothing, with some indication of a worse outcome. Which matters twice: the parts are tested rather than bundled on faith, and the most repeated advice about panic — control the breathing — is not where the work happens.

How it is run changes how well it works. Deacon and colleagues compared the procedure delivered along inhibitory-learning lines — varied, less predictable, not aimed at lowering anxiety within the session — against the conventional version, and the varied condition did better at follow-up. Craske's account explains why: the old association is not erased but out-competed, and the competitor has to be general enough to win outside the room it was built in.

Three separate reasons, rather than one.

Done for reassurance, it turns into a safety behaviour. A spin performed to confirm that nothing bad happens is a check, and checks need repeating. Whether an exercise is exposure or reassurance is settled by what the person is trying to find out, which is barely visible even from inside. The answer is procedural and it is most of the treatment: the prediction is stated in advance, out loud, in a form that can turn out false.

The design carries the effect and is not obvious. Which sensation to produce, matched to the one actually feared rather than the easiest to generate. Whether the usual precautions are still quietly running, in which case the exercise is being done with the insurance on. When to stop, which is not when anxiety reaches zero, because stopping at zero teaches that zero was the point.

Some of it is not safe for everyone. Over-breathing and straw breathing are physiological interventions, not metaphors. Cardiac conditions, respiratory disease and asthma, pregnancy, epilepsy, inner-ear disorders, uncontrolled high blood pressure and a history of fainting all change what can be attempted, and some rule parts of it out. That screen belongs to a clinician, and the person who most needs it is the one whose sensations have never been examined medically — which, by the order set out in panic attack, comes first anyway.

So this entry names the exercises and gives no procedure. The names are the useful part: they are how someone recognises what a therapist is proposing, and why being asked to make themselves dizzy on purpose is reasonable rather than a sign of being taken lightly. Worth knowing in advance: the first session is unpleasant, and anxiety rising before it falls is the normal course rather than a mistake.

What changes is narrower than people hope and more useful than it sounds. The sensations do not become pleasant and the body does not become less reactive. Dizziness becomes dizziness — an event with no sequel attached, caused by standing up too fast, uninteresting. The effect is subtraction of meaning rather than addition of tolerance, and it generalises because the catalogue of feared sensations is short while the situations that produce them are not.

The same component is used wherever the feared object is a bodily signal: in health anxiety beside work on body-checking, and at the breathless end of agoraphobia. But one element of a treatment is not a treatment, and nothing here is a plan. Chest pain, breathlessness, fainting or anything that felt different from the last time is a medical question first. Done properly the work takes weeks rather than years.

The panic attacks check scores the surge, the fear of the next one and what has already been given up separately, and the third column is where body-avoidance surfaces. The Anxiety Check reads the broader worry and the health anxiety test the version pointed at illness. All three are screens and say so.