Breathing and the autonomic nervous system: what the evidence says

Short answer

Slow-paced breathing reliably raises vagally mediated heart rate variability during and after practice [C1]. Meta-analyses of breathwork and HRV biofeedback find small to moderate average reductions in stress, anxiety and depression, mostly versus inactive controls [C5][C7][C8]. Brief breathing exercises for acute anxiety give inconsistent, technique-dependent results [C10].

What it is and how it shows up

The autonomic nervous system regulates heart rate, blood pressure and digestion without conscious effort. Breathing is unusual in being both automatic and voluntarily controllable, which makes it a practical lever on this system. Heart rate naturally rises slightly on inhalation and falls on exhalation; this is respiratory sinus arrhythmia, and heart rate variability (HRV) is the beat-to-beat variation that contains it [C3].

Slowing the breath, especially around the resonant frequency of the cardiovascular system (~ 0.1 Hz), increases cardiac oscillations and reflects improved vagally mediated heart rate variability and baroreflex sensitivity [C2]. The vagus nerve is the main parasympathetic output to the heart, and vagal afferents that project to interoceptive brain areas are stimulated more strongly during slow breathing than during spontaneous breathing [C2].

A systematic review and meta-analysis of 223 studies found that voluntary slow breathing increases vagally mediated HRV during the session, immediately after one session, and after multi-session practice [C1]. Slow breathing also shows central nervous system correlates: a systematic review of 15 articles found increased EEG alpha power and decreased theta power, alongside reduced symptoms of arousal, anxiety and depression in healthy subjects [C3].

In HRV biofeedback, a sensor displays the heart rhythm in real time while the person breathes slowly, usually at their own resonant frequency, so they can learn to keep the oscillations large [C2][C7]. The technique is typically practised for short sessions and used alongside other care rather than instead of it [C11].

In everyday terms, the family of techniques includes simply pacing the breath, paying attention to the breath without changing it, and paced breathing with biofeedback. They are often practised for a few minutes at a time, several times a day or week, and people usually turn to them for stress, anxiety or tension [C10]. The reviews below treat these variants separately, because their evidence differs [C10].

One measurement caution matters. Respiratory sinus arrhythmia and high-frequency HRV are usually read as markers of cardiac vagal activity, but the breathing pattern itself can change their amplitude without necessarily reflecting changes in tonic vagal activity, so these measures are only conclusive when interpreted relative to the breathing pattern [C4].

What research shows by approach

The evidence differs by technique, dose and comparison group, so this section walks through the main approaches one at a time.

Voluntary slow breathing and heart rate variability

The most consistent finding is physiological: across 223 studies, slow breathing increased vagally mediated HRV at every measured time point [C1]. This is a reliable short-term effect on a physiological index, not by itself proof of lasting clinical benefit [C1][C4]. The same review describes slow breathing as a low-tech, low-cost practice that can be used in prevention and alongside treatment, with few adverse effects expected [C11].

How the techniques differ

Breathing practices are not one intervention. Attention towards the breath and active control of the breath are different techniques, and in brief interventions for state anxiety they produced effects ranging from large to none depending on which was used [C10]. HRV biofeedback is the most structured variant: across 58 randomised trials it had its largest effects on anxiety, depression and anger, and its smallest on PTSD, sleep and quality of life [C7].

Breathwork for stress and mental health

Pooling 12 randomised controlled trials with 785 adults, breathwork was associated with significantly lower self-reported stress than control conditions, a small-to-medium effect (g = 0.35), with similar significant effects for anxiety (g = 0.32, 20 trials) and depressive symptoms (g = 0.40, 18 trials) [C5]. The authors urge caution against hype, since most studies were at moderate risk of bias [C5]. Because the outcome was self-reported stress, these numbers describe average group differences, not a fixed change for any individual [C5].

HRV biofeedback for stress and anxiety

HRV biofeedback adds real-time display of the heart rhythm to slow breathing. In 24 studies with 484 participants, training was associated with a large reduction in self-reported stress and anxiety (g = 0.81 pre-post, g = 0.83 versus controls) [C6]. More recent, larger syntheses are more conservative: across 58 randomised trials, HRV biofeedback showed a significant small to moderate effect that did not differ from other effective treatments, with larger effects against inactive than active controls [C7].

HRV biofeedback for depressive symptoms

In 14 RCTs with 794 adults, HRV biofeedback showed a medium mean effect on depressive symptoms (g = 0.38) [C8]. Remotely delivered HRV biofeedback across eighteen studies with 1352 participants showed a medium effect on depression (g = 0.41) and on HRV (g = 0.443), but the effect on stress was not significant [C9].

Brief breathing exercises for acute anxiety

For single-session relief of state anxiety, breathing-based techniques alone produced inconsistent results, with effects from large to none depending on the technique employed [C10]. Cognitive and embodiment techniques were more consistently helpful in the same review [C10].

Role as a complement

Slow breathing is described as a low-tech, low-cost technique with few adverse effects expected, and HRV biofeedback appears useful as a complementary treatment alongside other care [C11].

Limitations and uncertainty

Several limits matter. First, HRV indices are confounded by the breathing pattern itself, so physiological changes do not automatically prove changes in underlying vagal tone [C4]. Second, most meta-analytic effects are measured against inactive controls and rely on self-report; most breathwork trials were at moderate risk of bias, and effects shrink against active treatments [C5][C7]. Heterogeneity was substantial in several syntheses, and protocol and population characteristics changed the outcomes of remote biofeedback [C8][C9]. Third, brief breathing exercises for acute anxiety rest on few studies and give inconsistent results [C10]. Finally, adverse-event monitoring in breathing research is limited, so safety statements are expectations rather than measured rates [C11]. Pooled averages describe groups, not any one person's response. What would strengthen the evidence is more trials with active control conditions, longer follow-up, and systematic recording of unwanted effects; until then, the honest reading is that slow breathing reliably changes physiology and shows small to medium average symptom benefits, mostly against weak comparators [C5][C7].

When to see a professional

If anxiety, low mood or stress persist for weeks, interfere with work or relationships, or feel overwhelming, that is a reason to see a qualified clinician rather than relying on breathing exercises alone [C5][C10]. The same applies if breathing exercises themselves bring up distress or feel uncomfortable: techniques are studied as complements to care, not substitutes for it [C11]. A clinician can assess what is going on and discuss evidence-based options, which may include structured breathing or biofeedback practice where appropriate [C7][C11]. This page summarises research findings; it is not a diagnosis and not a treatment plan.

FAQ

What is slow breathing and how does it work?

Slow breathing means deliberately slowing the breath, often around the resonant frequency of the cardiovascular system (~ 0.1 Hz), which increases cardiac oscillations and vagally mediated heart rate variability [C2]. Across 223 studies it increased this index during, right after, and following multi-session practice [C1]. Reviews also report central nervous system correlates, such as more EEG alpha power and less theta power, and reduced feelings of arousal and anxiety in healthy adults [C3].

Do breathing exercises reduce stress and anxiety?

On average, by a small to medium amount. Breathwork across 12 trials with 785 adults showed a significant effect on self-reported stress (g = 0.35), with similar effects on anxiety and depressive symptoms [C5]. HRV biofeedback showed a large pre-post reduction in stress and anxiety in an older meta-analysis of 24 studies [C6].

Does HRV biofeedback work better than plain breathing?

Not clearly. Across 58 randomised trials HRV biofeedback had a significant small to moderate effect that did not differ from other effective treatments, and effects were larger against inactive than active controls [C7]. For depression specifically, 14 RCTs with 794 participants gave a medium effect (g = 0.38) [C8].

Can breathing exercises help depression?

Possibly, as an add-on. HRV biofeedback showed a medium effect on depressive symptoms in 14 RCTs (g = 0.38) [C8], and remote biofeedback across eighteen studies showed a medium effect on depression (g = 0.41) though not on stress [C9]. These are complementary options, not replacements for clinical care [C11].

Are breathing exercises safe?

They are generally described as low-cost with few adverse effects expected [C11]. One caveat: HRV measures used in biofeedback are influenced by the breathing pattern itself, so tracking them without context can mislead [C4]. If a practice consistently makes you feel worse, mention it to a clinician [C11]. Researchers note that unwanted effects are not always recorded systematically, so the true rate of problems is uncertain [C11].

When should I talk to a professional about breathing exercises?

If anxiety or low mood persist, worsen, or interfere with daily life, contact a qualified clinician rather than managing it alone [C10][C11].

Self-help information does not replace a clinician; people with severe or persistent symptoms should contact a professional or local emergency services.