Comparison of diaphragm EMG centroid frequencies: esophageal versus chest surface leads

Previous work has suggested that diaphragm EMGs recorded from the lower thoracic wall yield spuriously low centroid frequencies (Fc). For this reason we compared EMGs recorded from two anterolateral thoracic wall locations with EMGs recorded simultaneously from a bipolar esophageal electrode in 11 stable quadriplegic patients. Their maximal inspiratory pressures (Pdimax) ranged from 31 to 85, averaging 55 cm H2O. Quadriplegic patients were selected to exclude EMG contamination of surface recordings by intercostal muscle activity. Recordings were made during resting breathing, voluntary hyperventilation, CO2-stimulated hyperventilation, and inspiratory resistance breathing at 50% of Pdimax. Centroid frequencies were consistently lower from surface recordings than from esophageal recordings. Centroid frequencies (Hz) recorded from the esophagus were 86.9 +/- 3.0 (SEM) resting, 93.4 +/- 3.7 during voluntary hyperventilation, 85.8 +/- 4.2 during CO2-stimulated ventilation, and 88.5 +/- 1.4 during inspiratory resistance breathing. Corresponding Fc values from simultaneous thoracic surface recordings were 62.6 +/- 3.6, 62.0 +/- 2.8, 58.6 +/- 3.1, and 58.8 +/- 2.5. On several subjects, decreases in Fc that occurred during sustained maximal inspiratory efforts were present in esophageal recordings but not in surface recordings. We conclude that thoracic surface recordings of the diaphragmatic EMG do not accurately reflect frequency information.