Does breathwork improve stress and mental health through interoceptive mechanisms, and what do randomized trials show about specificity?

What randomized trials actually show about breathwork and distress

Across randomized controlled trials, breathwork consistently produces small-to-moderate improvements in self-reported distress. A meta-analysis of 12 trials with 785 adults found that breathwork lowered subjective stress more than non-breathwork controls, with a small-to-medium effect size (g = −0.35); similar effects appeared for anxiety (g = −0.32 across 20 trials) and depressive symptoms (g = −0.40 across 18 trials) [1]. A separate systematic review of 15 randomized trials described moderate-to-strong support for breathwork and chanting in relieving anxiety, depression, stress, and PTSD symptoms, though study quality varied [2].

The size of these effects matters for interpretation. An effect size around 0.3 to 0.4 is meaningful but not dramatic—roughly the kind of shift you might see from a modest but real intervention, not a transformation. The meta-analysis authors themselves urged caution, noting most trials carried moderate risk of bias and calling for better-designed studies to avoid a mismatch between public enthusiasm and evidence [1]. In a more recent single-blind trial of 98 paramedicine students, a 12-week slow-breathing protocol (5-second inhale, 5-second exhale, 2-second hold, practiced twice daily) significantly reduced stress, anxiety, and depression and raised resilience compared with controls, with medium-to-large effects; insomnia and general wellbeing did not change significantly [5]. That trial is notable because it tested a structured protocol in a high-stress group during exam periods, but it is one study in one population.

Is interoception the mechanism, or just a plausible story?

Interoception refers to sensing, interpreting, and regulating internal body signals—heartbeat, breathing, gut sensations. Slow breathing is an obvious candidate for training this capacity, since it deliberately manipulates an internal signal and asks you to notice it. But the randomized evidence connecting breathwork to mental health through interoception is thin. The breathwork meta-analysis and the chanting/breathwork review both measured symptom outcomes and did not examine interoceptive mechanisms at all [1][2].

The closest direct test is a subclinical randomized trial of a program called InMe, which combined slow breathing with training intended to enhance interoceptive insight and confidence in reading body signals under stress [4]. Its design targets exactly the proposed pathway—breath practice building interoceptive self-efficacy, which in turn reduces mental health symptoms—but the available report is a partial abstract, so the size and durability of any mediated effect cannot be judged from it. Separately, a systematic review of 31 randomized trials of interoception-based interventions found that 20 (about 65%) improved interoception more than control conditions, with the most promising results for PTSD, irritable bowel syndrome, fibromyalgia, and substance use disorders—yet evidence that these changes translated into symptom improvement was inconclusive [3]. That gap is important: even when interoception improves, symptom relief does not reliably follow, which weakens the case that interoception is the main route by which breathwork helps.

Does it matter which breathing practice you do, or who you are?

The trials do not answer the specificity question cleanly. The meta-analysis compared breathwork with non-breathwork controls, which shows breathwork beats doing nothing or an unrelated activity, but not that one breathing pattern outperforms another or outperforms relaxation, attention control, or expectancy alone [1]. The paramedicine trial used a specific paced protocol and found benefits for stress, anxiety, depression, and resilience but not insomnia or psychological wellbeing [5]—a pattern suggesting effects may be outcome-specific rather than global, though a single trial cannot establish that.

Population also shapes what the evidence means. The meta-analysis pooled mostly adult samples with moderate risk of bias [1]; the paramedicine trial studied students under exam stress [5]; the InMe trial was subclinical [4]. None of these establishes benefit for clinical anxiety or depressive disorders, and the interoception review found symptom improvement inconclusive even in diagnosed populations [3]. A reasonable reading is that breathwork is a low-cost self-regulation practice with modest average benefits for self-reported distress, while claims that it works specifically by retraining interoception, or that a particular protocol is superior, outrun what randomized trials have shown.

Sources used in this answer

Related research questions