Why "take a deep breath" is the wrong advice for panic.

The advice on every poster

“Take a deep breath” may be the most repeated piece of calming advice in the English language. It is on posters, in movies, in the mouths of well-meaning friends, and inside most wellness apps. It survives because it sounds physiological and because sometimes, for mild stress, it is harmless.

For panic it is not harmless. The instruction misreads what the body is doing during a panic attack, and following it enthusiastically can make the attack worse. This is not a contrarian opinion. It has been a documented concern in the clinical breathing literature for over two decades.

What panic does to the breath

During a panic attack, breathing typically becomes fast and shallow, and often deep as well. The person is moving more air than their body needs. The result is hyperventilation: carbon dioxide is exhaled faster than the body produces it, and blood CO2 drops below its normal range, a state called hypocapnia.

Low CO2 is not neutral. It constricts blood vessels, including in the brain, and shifts blood chemistry in ways that produce a very specific symptom cluster: light-headedness, tingling in the hands and face, chest tightness, a feeling of not getting enough air. Read that list again. It is the panic attack symptom list. A large part of what makes panic feel like dying is the direct physiology of over-breathing.

This creates the loop. The person feels short of air, so they breathe harder. Breathing harder lowers CO2 further, so the symptoms intensify, which confirms that something is badly wrong, which drives the breathing harder still. “Take a deep breath” is an instruction to tighten that loop.

A large part of what makes panic feel like dying is the direct physiology of over-breathing. Deeper breaths tighten the loop.

What the literature actually says

In 2003, Alicia Meuret and colleagues published a paper with an unusually pointed title for a clinical journal: “Breathing training for treating panic disorder: useful intervention or impediment?” Their argument, built on the hyperventilation research, was that breathing instructions which encourage deeper breathing can act against the patient, and that effective breathing training for panic should target slower, shallower breathing and normalized CO2.

Meuret’s later work went further, using CO2 biofeedback to teach panic patients to raise their CO2 levels by breathing less, not more, with significant reductions in panic symptoms. The direction of the correction is the opposite of the poster advice.

None of this means breathing techniques are useless for panic. It means the technique has to match the physiology: slow the breath down, keep it small, and put the emphasis on the exhale, the phase of the breath that engages the parasympathetic brake on the heart.

The cue that works instead

The corrected instruction fits in one line: small, slow breaths through the nose, with the exhale longer than the inhale.

The best-evidenced structured version of that idea is cyclic sighing, two short inhales through the nose followed by one long, slow exhale through the mouth. The double inhale re-opens the small air sacs of the lungs without moving a large volume of air, and the extended exhale does the calming work. In a 2023 Stanford randomized controlled trial, five minutes a day of cyclic sighing improved positive affect more than mindfulness meditation and reduced respiratory rate more than the other techniques studied, box breathing included.

It also has the lowest cognitive load of any structured technique. There is nothing to count and nothing to hold. In a panic attack, that is not a convenience. It is the difference between a technique that gets used and one that gets abandoned at the exact moment it was needed.

• · Small, slow breaths through the nose. Not deep ones.
• · Exhale longer than you inhale, every cycle.
• · Cyclic sighing: two short inhales, one long exhale through the mouth.
• · No counting, no breath-holds, nothing to remember.

Breathwork has fine print

The other thing the poster never mentions: breathing techniques are not uniformly safe for everyone.

Breath-retention practices (long holds, hyperventilation-style rounds) can be genuinely destabilizing for panic-prone people, whose alarm systems are exquisitely sensitive to CO2 shifts. A breath hold can literally trigger the alarm it was supposed to quiet.

Pregnancy changes the calculus too. Sustained holds and forceful patterns are generally not what a pregnant body needs. And practices like mouth-taping, popular in wellness content, have no place in a tool used by people with anxiety or sleep-disordered breathing.

The wellness industry mostly ships breathing content the way it ships playlists: same catalog for everyone, no questions asked. Clinically, that is negligent-by-default. The right pattern for a given nervous system depends on who is breathing.

What we built with this

The phrase "take a deep breath" does not appear anywhere in Stillee's panic flow. That is a rule, not an accident. The flow opens with cyclic sighing already in motion, the visual leads the rhythm so there is nothing to count, and the written cues follow the clinical direction: small breaths, long exhale.

The wider breathing library is organized by what techniques do, not by what a manual calls them: the Circuit Breaker for the acute downshift, the Daily Practice, the Wind-Down for pre-sleep, the Activator for the low-energy slump. Cyclic sighing is the hero of the set: it is the same pattern from the Stanford trial, and the same one that anchors the panic flow, so it is one move you learn once and meet everywhere.

Every session is haptic-guided (the pacing pulses gently in your hand) so it works with the screen barely glanced at, or in the dark, including a 3 AM breathing mode with a pitch-black screen and no clock staring back at you.

The contraindications are built in, not bolted on. Stillee screens techniques against your profile: if you have told us about panic, we will not hand you a breath-hold. Pregnancy-safe defaults for pregnant users. No mouth-taping content anywhere.

A closing caution in both directions. If slow-exhale breathing reliably fails to help, or panic attacks are frequent, that is a conversation for a clinician, and cognitive behavioral therapy for panic is the best-evidenced path. And if a technique anywhere, in any app, asks a panicking person to breathe deeper, it is safe to skip.