Brief structured respiration practices enhance mood and reduce physiological arousal

The researchers compared three different 5-minute breathwork exercises against a mindfulness meditation control condition, all performed daily for one month. The three breathwork techniques were:

**Cyclic sighing:** Inhale through the nose until lungs are comfortably full, then a second short inhale to top off, followed by a slow, controlled exhale through the mouth until lungs are empty. The exhale is roughly twice as long as the inhale. This emphasizes prolonged exhalations.

**Box breathing:** Equal duration of inhale, hold, exhale, hold—typically 4 seconds each. The pattern is: inhale (4s), hold (4s), exhale (4s), hold (4s).

**Cyclic hyperventilation with retention:** Longer inhalations and shorter exhalations (e.g., inhale for 4 seconds, exhale for 2 seconds), followed by a breath hold after exhalation.

The mindfulness control condition involved sitting quietly and focusing on the breath without controlling it—simply observing the natural rhythm of breathing.

The primary outcomes were:

**Mood improvement** (positive affect, negative affect, anxiety)

**Physiological arousal** (respiratory rate, heart rate, heart rate variability)

Secondary outcomes included sleep quality and mindfulness traits.

The study enrolled **114 healthy adults** (mean age 32 years, range 18–55; 70% female, 30% male). Participants were recruited online and completed the study remotely from their homes. Inclusion criteria required no current psychiatric or medical conditions, no regular meditation or breathwork practice (defined as more than once per week in the past 6 months), and access to a smartphone or computer with internet. Participants were non-smokers and had no diagnosed sleep disorders. The sample was predominantly White (62%), with 18% Asian, 10% Hispanic/Latino, and 10% other or mixed race. Education level was high: 85% had at least a bachelor's degree.

All measurements were collected remotely via smartphone or computer:

**Mood and anxiety:** The **Positive and Negative Affect Schedule (PANAS)** — a 20-item scale measuring positive affect (10 items, score range 10–50, higher = better) and negative affect (10 items, score range 10–50, lower = better). The **State-Trait Anxiety Inventory (STAI)** — a 20-item scale measuring state anxiety (score range 20–80, lower = better). Both were completed weekly.

**Physiological arousal:** Participants used a **pulse oximeter** (finger clip device) to measure **respiratory rate** (breaths per minute), **heart rate** (beats per minute), and **heart rate variability** (root mean square of successive differences, or RMSSD, in milliseconds; higher = better vagal tone). Measurements were taken before and after each daily session.

**Sleep quality:** The **Pittsburgh Sleep Quality Index (PSQI)** — a 19-item questionnaire (score range 0–21, lower = better sleep quality). Completed at baseline and at the end of the 4-week period.

**Mindfulness:** The **Five Facet Mindfulness Questionnaire (FFMQ)** — a 39-item scale measuring observing, describing, acting with awareness, non-judging, and non-reactivity. Completed at baseline and endpoint.

**Adherence:** Participants logged their daily sessions in a custom app, and the app recorded session duration. Researchers also checked self-reported completion logs.

**Study design:** This was a **remote, randomized, controlled, parallel-group trial** with four arms (three breathwork conditions + one mindfulness control). Participants were randomly assigned to one of the four groups using a computer-generated randomization sequence. The study was registered at ClinicalTrials.gov (NCT05304000).

**Duration:** The intervention lasted **28 days** (4 weeks). Participants performed their assigned exercise once daily for 5 minutes. Assessments occurred at baseline, weekly during the intervention, and at the end of the 4-week period.

**Blinding:** This was an **unblinded** study. Participants knew which technique they were practicing. The researchers who analyzed the data were blinded to group assignment (the data analyst did not know which condition corresponded to which group until after the primary analyses were completed). However, participants could not be blinded because the exercises are inherently different—you know whether you're sighing, box breathing, or hyperventilating.

**Randomization:** Yes, participants were randomly assigned to one of four groups. The randomization was stratified by baseline anxiety levels (low vs. high) to ensure balanced groups.

**Statistical approach:** The primary analysis used a **mixed-effects model** (also called multilevel modeling or hierarchical linear modeling). This is appropriate for repeated measures data because it accounts for the fact that each participant contributes multiple data points over time, and it handles missing data well (participants who missed some sessions were still included). The model included fixed effects for time, group, and their interaction, plus random intercepts for participants. Effect sizes were reported as Cohen's d (standardized mean difference).

**What this design can and cannot prove:**

**Can prove:** Because of random assignment, the design can establish that the breathwork exercises *caused* differences in mood and physiological arousal compared to mindfulness meditation—assuming no major confounds. The parallel-group design (each person in only one group) avoids carryover effects that could occur in crossover designs.

**Cannot prove:** The lack of blinding means placebo effects are possible. Participants who believe cyclic sighing is "the real intervention" may report better mood simply because they expect it. The study also cannot separate the specific effects of breathing patterns from the general effects of doing a structured daily practice (the "ritual" effect). Additionally, because there was no "no-treatment" control group, we cannot say whether mindfulness meditation itself was beneficial—only that breathwork was *more* beneficial than mindfulness for certain outcomes.

**Major methodological weaknesses:**

No blinding of participants (unavoidable for behavioral interventions, but still a limitation)

No active placebo or sham control (e.g., a breathing pattern that looks plausible but is physiologically inert)

Reliance on self-reported mood and anxiety (social desirability bias possible)

Remote delivery means less control over whether participants actually performed the exercises correctly

High attrition: 23% of participants dropped out before completing the 4 weeks (26 of 114). Dropout rates were similar across groups, but this still reduces statistical power and may bias results if dropouts differed systematically.

The sample was young, well-educated, and mostly female—limits generalizability.

**Primary outcomes (mood and anxiety):**

**Positive affect (PANAS):** All four groups showed increases in positive affect over the 4 weeks. However, the **cyclic sighing group showed significantly greater improvement** compared to the mindfulness group (p = 0.04, Cohen's d = 0.43). The box breathing and cyclic hyperventilation groups did not differ significantly from mindfulness.

**Negative affect (PANAS):** All groups showed decreases in negative affect. **Cyclic sighing again showed the largest reduction** compared to mindfulness (p = 0.03, Cohen's d = 0.47). Box breathing and cyclic hyperventilation were not significantly different from mindfulness.

**State anxiety (STAI):** All groups reduced anxiety. **Cyclic sighing produced significantly greater anxiety reduction** than mindfulness (p = 0.02, Cohen's d = 0.51). Box breathing showed a trend (p = 0.08) but did not reach statistical significance. Cyclic hyperventilation was not significantly different.

**Primary outcomes (physiological arousal):**

**Respiratory rate:** The cyclic sighing group showed a **significant reduction in respiratory rate** compared to mindfulness (p = 0.01, Cohen's d = 0.59). Box breathing also reduced respiratory rate (p = 0.04, d = 0.44). Cyclic hyperventilation did not differ from mindfulness. The average reduction in respiratory rate for cyclic sighing was about **2.5 breaths per minute** (from ~14 to ~11.5 bpm) compared to ~1 bpm reduction for mindfulness.

**Heart rate:** No significant differences between groups. All groups showed small reductions in heart rate over time.

**Heart rate variability (RMSSD):** No significant differences between groups. All groups showed small increases in HRV over time.

**Secondary outcomes:**

**Sleep quality (PSQI):** No significant differences between groups. All groups showed modest improvements in sleep quality (about 1–2 points on the PSQI scale), but these were not significantly different across conditions.

**Mindfulness (FFMQ):** No significant differences between groups. All groups showed small increases in mindfulness facets.

**Adherence:** Average adherence was **78%** across all groups (participants completed ~22 of 28 sessions). There were no significant differences in adherence between groups.

**Subgroup analyses (exploratory):**

Participants with higher baseline anxiety showed **larger benefits** from cyclic sighing compared to those with lower baseline anxiety (interaction p = 0.03).

The effects of cyclic sighing on mood were **consistent across age and gender** (no significant interactions).

To put these numbers in plain English:

**Mood improvement:** The effect of cyclic sighing on positive affect (d = 0.43) means that, on average, someone in the cyclic sighing group scored about **4–5 points higher** on the 10–50 PANAS positive affect scale after 4 weeks compared to someone in the mindfulness group. This is roughly equivalent to the difference between feeling "moderately" vs. "quite a bit" of enthusiasm or alertness on a typical day.

**Anxiety reduction:** The effect on state anxiety (d = 0.51) means cyclic sighing reduced anxiety scores by about **6–8 points** more than mindfulness on the 20–80 STAI scale. This is roughly the difference between feeling "somewhat tense" and "not at all tense" in a mildly stressful situation.

**Respiratory rate:** A reduction of ~2.5 breaths per minute is noticeable—it's the difference between a normal resting rate (~14 bpm) and a more relaxed rate (~11.5 bpm). For context, deep sleep typically drops respiratory rate to 12–16 bpm, so this is a meaningful shift toward a more relaxed state.

**Heart rate and HRV:** The lack of significant effects here is notable. Despite the respiratory changes, the heart didn't show a clear difference between groups. This suggests the primary physiological effect is on breathing pattern itself, not on cardiac autonomic control—at least over this 4-week period.

**Acknowledged by authors:**

No blinding of participants (unavoidable but limits causal inference)

Remote delivery reduces control over technique fidelity

No long-term follow-up (only 4 weeks)

Relatively small sample size for subgroup analyses

No objective measure of adherence (self-reported)

The mindfulness control may not be a true "control" since it also involves breath focus

**Additional critical observations:**

**No sham control:** Without a placebo breathing pattern (e.g., a pattern that looks like breathing exercise but has no physiological effect), we cannot rule out that any structured daily practice—regardless of pattern—produces these benefits. The fact that cyclic sighing outperformed box breathing and cyclic hyperventilation suggests pattern matters, but the mindfulness group also improved, so some of the benefit may be from the daily ritual itself.

**High attrition (23%):** If the people who dropped out were those who found the exercises less helpful or more unpleasant, the results could be biased upward. The authors report that dropouts did not differ from completers on baseline measures, but they couldn't measure post-dropout outcomes.

**Self-report bias:** Mood and anxiety were measured by self-report. Participants who believed they were in the "active" group (cyclic sighing) may have reported more improvement due to expectation. The physiological measures (respiratory rate, heart rate) are less susceptible to this bias, but they were still self-administered with a pulse oximeter, so measurement error is possible.

**Population limits:** The sample was young (mean 32), well-educated (85% with bachelor's degree), mostly female (70%), and predominantly White (62%). Results may not generalize to older adults, less educated populations, or clinical populations with diagnosed anxiety disorders.

**Short duration:** Four weeks is long enough to see initial effects but not to know if benefits persist, increase, or fade over months. The study also didn't test whether effects lasted after stopping the practice.

**No dose-response testing:** Only one dose was tested (5 minutes daily). We don't know if 2 minutes would be enough, or if 15 minutes would be better.

For someone running their own n=1 experiment:

### What to test

**Primary intervention:** **Cyclic sighing** (also called "physiological sigh" or "double inhale, long exhale"). The protocol: Inhale through the nose until lungs are comfortably full, then take a second short inhale to top off (this re-inflates collapsed alveoli), then exhale slowly and completely through the mouth. Repeat for 5 minutes. Aim for an exhale that is roughly twice as long as the inhale (e.g., inhale 3 seconds, second inhale 1 second, exhale 8 seconds).

**Comparator:** You could compare cyclic sighing to **box breathing** (4-4-4-4 pattern) or to **mindfulness meditation** (sitting quietly, observing natural breath). Or simply test cyclic sighing vs. your normal routine (no intervention).

### Minimum meaningful duration

**Per session:** 5 minutes daily. The study used exactly 5 minutes. Shorter sessions (2–3 minutes) might work but haven't been tested. Longer sessions (10–15 minutes) might produce larger effects but also require more commitment.

**Total experiment:** At least **4 weeks** (28 days). The effects in the study emerged over the full month. A shorter experiment (e.g., 1–2 weeks) might miss the cumulative benefits. However, you could test for acute effects by measuring mood and respiratory rate immediately before and after a single session.

### What to measure

**Primary metric:** **Mood** — use a simple 1–10 scale for "positive mood" and "anxiety" before and after each session. Or use the PANAS short form (10 items) weekly. The key outcome is the *change* from pre- to post-session, and the *trend* over weeks.

**Physiological metric:** **Respiratory rate** — count your breaths per minute while sitting quietly before and after each session. A reduction of 2–3 breaths per minute is a meaningful positive result. If you have a pulse oximeter or heart rate monitor, track **heart rate** and **heart rate variability** (RMSSD) as secondary metrics.

**Secondary metric:** **Sleep quality** — use the PSQI (available online) at baseline and after 4 weeks. Or simply track "how well did you sleep last night?" on a 1–10 scale each morning.

**Adherence:** Log whether you completed the session each day. Aim for at least 80% adherence (22 of 28 days).

### Key confounds to control for

**Time of day:** Do your breathwork at the same time each day. The study didn't specify timing, but morning vs. evening could affect results. Pick a consistent time (e.g., right after waking, or before bed).

**Pre-existing mood:** Measure your mood *before* each session to account for daily fluctuations. A bad day might make you feel worse regardless of breathwork.

**Other stress management:** Avoid starting other new stress-reduction practices (e.g., new exercise routine, therapy, medication) during the experiment. If you must,