Table 3 Summary of sEMG parameters and their applications and limitations

Magnitude of muscle activity

Amplitude

Difference between maximum and minimum value during. one breath (either including or excluding the baseline)

Using the 95th and 5th percentile to calculate this difference may be more robust

Assessing changes in absolute magnitude of muscle activity within a single recording

Low amplitude does not imply low muscle activity and vice versa

Only comparable within short-time recordings

Does not enable between-patient or between-recording comparisons

Amplitude normalized to maximum breathing effort

Amplitude divided by maximum amplitude obtained during maximum inspiratory maneuver

Assessing changes in relative muscle activity

Improves sEMG amplitude interpretability

Maximum inspiratory maneuvers could be challenging to perform in critically ill patients and multiple repetitions are required

Maximum amplitude should be re-obtained for a new recording

Amplitude normalized to maximum amplitude within recording

Amplitude divided by maximum amplitude obtained over a given measurement (without ensuring maximum effort)

Assessing changes in relative muscle activity within a patient during a recording

Improves sEMG amplitude interpretability

Maximum amplitude should be re-obtained for a new recording

It does not enable between-patient comparisons or within-patient comparisons across multiple recordings

EMG-time product

Area under the sEMG envelope, per breath or per time unit

Less sensitive to remaining artifacts than computing breathwise sEMG amplitudes

Dependent on whether the baseline is included in computation

Affected by sEMG onset and offset definitions

Estimation of mechanical output

Estimated breathing effort

Pmus = k x sEMG, with conversion factor k obtained from patient-specific measures (end-expiratory occlusion or model-based)

Translates muscle activity to mechanical output

k needs to be re-evaluated for a new recording

Assumes a linear relationship between muscle activity and output

Timing of muscle activity

Time-to-peak

Time from onset to peak sEMG

Suggested to reflect respiratory drive

Onset/offset is not binary; no clear definition exists (see text for approaches)

Increase in sEMG activity may not be linear

Unclear comparability between patients

Duration of muscle activity

Time from sEMG onset to offset

Informs about the duration of muscle activation

Onset/offset is not binary; no clear definition exists (see text for approaches)

Phase angle

Phase lag between sEMG onset (or offset) and start (or end) of ventilator pressurization

Phase lag between onset (or offset) of multiple sEMG signals

Assessing patient-ventilator interaction

Assessing activation patterns (different muscles)

Onset/offset of sEMG is not binary; no clear definition exists (see text for approaches)

Absolute time delay

Time delay (in ms) between sEMG onset (or offset) and start (or end) of ventilator pressurization

Time delay (in ms) between onset (or offset) of multiple sEMG signals

Assessing patient-ventilator interaction

Assessing activation patterns (different muscles)

Onset/offset of sEMG is not binary; no clear definition exists (see text for approaches)

Does not correct for duration of activity/pressurization such as with phase angle

Fatigue

Fatigue onset

Various metrics have been described:

Shift in mean (or median) frequency

High/low frequency ratio (H/L ratio, with H = 150–350 Hz and L = 20–46.7 Hz)

Spectral moments ratio of order five (SMR5) and fuzzy approximate entropy (fApEn)

May inform about diaphragm fatigue before a decrease in pressure-generating capacity occurs

No data and cutoff values exist

Challenging to compute reliably in respiratory sEMG due to low SNR