Comparing ECG artifact removal methods for diaphragm EMG during inspiratory pressure-threshold loading: Effects on signal quality and the EMG-pressure-time product relationship.

Diaphragm electromyogram (EMGdi) is commonly measured using esophageal multi-paired electrode catheters; however, contamination by cardiac electrical activity can compromise signal quality. This study compared several approaches for reducing electrocardiographic (ECG) artifact and examined whether their performance differed based on sex. We also evaluated their influence on the relationship between EMGdi and force output (PTPdi). Sixteen healthy young adults (7 female, 9 male) performed inspiratory pressure threshold loading (PTL) to task failure while instrumented with a multi-pair esophageal balloon catheter and a simultaneous ECG. EMGdi signals were processed using four approaches: (1) selection of EMG segments between QRS complexes (ECG-R), (2) independent component analysis (ICA), (3) ICA with wavelet transformation (ICA-W), and (4) no artifact removal (RAW). Estimated marginal slopes relating EMGdi to inspiratory PTL did not differ between females and males or among processing methods (all p > 0.05). EMGdi was significantly associated with PTPdi for all processing methods in both sexes. In females, however, these associations were weaker following ICA and ICA-W than with ECG-R (p < 0.05), whereas no differences between methods were observed in males. These findings indicate that all methods effectively reduce ECG artifact during PTL when heart rate remains relatively low. When a simultaneous ECG is unavailable, ICA and ICA-W provide suitable alternatives. However, for studies examining the relationship between EMGdi and diaphragm force output, ECG-guided QRS exclusion is the preferred processing approach.