Table 3.
Analysis of breathing during initial experimental conditions.
| Dependent variable | Comparison | Estimate | SE | z | p-value |
|---|---|---|---|---|---|
| (ln-transformed) | R vs. U | 0.27 | 0.03 | 11.17 | < 0.0001 |
| vs. R | − 0.14 | 0.03 | − 4.58 | < 0.0001 | |
| FRC | R vs. U | − 0.35 | 0.01 | − 32.67 | < 0.0001 |
| vs. R | 0.18 | 0.02 | 10.38 | < 0.0001 |
The mixed-effects regression models used to test these contrasts accounted for a subject’s age, their BMI, and the experimental phase in the fixed effects, and included a random effect of subject identity to account for repeated measures (n = 911 breaths from 17 subjects in phases R, U, and ). To appropriately model the variances, was ln-transformed, and the analysis for FRC accounted for unequal variance across phases. The table provides test statistics for post-hoc contrasts of specific comparisons. U = initial standing upright measurement; R = unweighted prone measurement; = first minute of weighted prone measurement.