Comparing the effects of square, 4-7-8, and 6 breaths-per-minute breathing techniques on heart rate variability, CO2 levels, and affect.
Josh Marchant, B.S., Mikel Cressman, M.S.
Mentored by: Patrick Steffen, Ph.D. and Inna Khazan, Ph.D.
And a thank you to our wonderful RAs!
5:5
5-second inhale
5-second exhale
4:6
4-second inhale
6-second exhale
Square
4-second inhale
4-second hold
4-second exhale
4-second hold
4-7-8
4-second inhale
7-second hold
8-second exhale
Four Breathing Techniques
Evidence for Square/Box Breathing?
Evidence for 4-7-8 Breathing?
Evidence for 6 breaths per minute (5:5 and 4:6) Breathing
Heart Rate Variability (HRV)
End Tidal CO2 (PETCO2)
Blood Pressure
Affect
- Scale of Positive and Negative Experience (SPANE)
- Has demonstrated good reliability, convergent validity, and divergent validity with other measures of mood/affect (Jovanović et al., 2020).
Design
Design
Descriptive statistics for demographic variables.
| N (%) | M | SD | Range |
Total sample size | 84 | — | — | — |
Age | — | 20 | 1.76 | 17 - 25 |
Sex | — | — | — | — |
Female | 50 (59.5) | — | — | — |
Male | 34 (40.5) | — | — | — |
Race | — | — | — | — |
White | 72 (85.7) | — | — | — |
African American | 0 (0) | — | — | — |
Asian/Pacific Islander | 5 (6) | — | — | — |
Mixed Race | 7 (8) | — | — | — |
Ethnicity | — | — | — | — |
Hispanic | 10 (11.9) | — | — | — |
Not Hispanic | 74 (88) | — | — | — |
BMI | — | 23.33 | 3.98 | 16.3 - 36.9 |
Percentage adherence by breathing technique.*
Technique | Target breathing rate (BPM) | Percent Adherence (+/- 1 second) | Percent Adherence (+/- 2 seconds) |
5:5 | 6.00 | 91.7% | 96.4% |
4:6 | 6.00 | 85.7% | 95.2% |
Square | 5.00 | 59.5% | 94.0% |
4-7-8 | 3.16 | 72.6% | 88.1% |
*ICCs for counting breathing rate with respiration belt were 0.87 (5:5), 0.75 (4:6), 0.80 (square), and 0.74 (4-7-8).
Hypothesis 1
| | 95% LL | 95% UL | p | Scheffé adjusted p |
RMSSD change from baseline |
|
|
|
|
|
5:5 | 16.04 | 11.07 | 21.00 | >0.001 | >0.001 |
4:6 | 22.64 | 17.68 | 27.61 | >0.001 | >0.001 |
Square | 7.45 | 2.48 | 12.42 | 0.003 | 0.071 |
4-7-8 | 7.59 | 2.62 | 12.55 | 0.003 | 0.063 |
RMSSD pairwise comparisons |
|
|
|
|
|
5:5 vs. square | 8.59 | 0.77 | 16.41 |
| 0.022 |
5:5 vs. 4-7-8 | 8.45 | 0.63 | 16.27 |
| 0.026 |
4:6 vs. square | 15.20 | 7.38 | 23.01 |
| >0.001 |
4:6 vs. 4-7-8 | 15.06 | 7.24 | 22.88 |
| >0.001 |
4:6 vs. 5:5 | 6.61 | -1.21 | 14.43 | | 0.146 |
LF log pairwise comparisons |
|
|
|
|
|
5:5 vs. square | 0.25 | 0.03 | 0.47 |
| 0.019 |
5:5 vs. 4-7-8 | 0.23 | -0.01 | 0.45 |
| 0.041 |
4:6 vs. square | 0.29 | 0.07 | 0.51 |
| 0.003 |
4:6 vs. 4-7-8 | 0.27 | 0.05 | 0.49 |
| 0.007 |
*Counterbalanced group was not associated with outcomes
LF log pairwise comparisons | | 95% LL | 95% UL | Scheffé adjusted p |
Square vs. 5:5 | 1.35 | 1.07 | 1.63 | >0.001 |
4-7-8 vs. 5:5 | 1.61 | 1.33 | 1.90 | >0.001 |
Square vs. 4:6 | 1.17 | 0.88 | 1.45 | >0.001 |
4-7-8 vs. 4:6 | 1.43 | 1.14 | 1.71 | >0.001 |
Potentially unhealthy
*VLF is not as accurate during 5-minute measurement segments
Hypothesis 2
Hypothesis 3
PETCO2 change from baseline | | 95% LL | 95% UL | p | Scheffé adjusted p |
5:5 | -3.87 | -5.09 | -2.65 | >0.001 | >0.001 |
4:6 | -4.27 | -5.50 | -3.05 | >0.001 | >0.001 |
Square | 0.28 | -1.19 | 1.25 | 0.964 | 1.000 |
4-7-8 | 1.17 | -0.05 | 2.39 | 0.060 | 0.468 |
PETCO2 pairwise comparisons |
|
|
|
|
|
5:5 vs. square | -3.90 | -5.83 | -1.97 |
| >0.001 |
5:5 vs. 4-7-8 | -5.04 | -6.97 | -3.12 |
| >0.001 |
4:6 vs. square | -4.30 | -6.23 | -2.37 |
| >0.001 |
4:6 vs. 4-7-8 | -5.44 | -7.38 | -3.51 |
| >0.001 |
- Machine was not zeroed before some CO2 measurements were taken, leading to some inaccurate baseline results. However, change scores were accurate.
- Excluding participants whose baseline was not in the normal (35-45 range) results in nearly the same Beta and p-values.
Hypothesis 4
Conclusions
Exploratory – email sent after study completion; only half of the sample responded
Limitations
Main Takeaway
References
References