Reliability of digital filtering on surface electromyographic activity of respiratory muscles in patients with COVID-19

Abstract

Introduction: Surface electromyography (sEMG) is valuable for assessing respiratory muscle activity, but the influence of electromagnetic fields can impair the accuracy of the signals. This study aims to compare the Root Mean Square (RMS) values of sEMG signals from respiratory muscles in COVID-19 patients receiving oxygen therapy in the ICU before and after digital filtering. Methods: This exploratory analysis is part of a cross-sectional study registered on ClinicalTrials.gov (NCT05572853). sEMG signals from the sternocleidomastoid, scalenus, diaphragm, and rectus abdominis muscles were recorded in COVID-19 patients receiving low-flow oxygen therapy. The RMS data were analyzed and processed using Scilab 6.1.1 software before and after applying a filter with a 60 Hz notch and a passband between 20-500 Hz. Results: 71 patients with a mean age of 55 ± 14 years, predominantly male (73.2%), were included. Muscle’s activity remained unchanged after signal filtering, showing higher RMS intensity (pre-filter/post-filter) for the scalene (14.0 ± 8.6 / 8.5 ± 6.2), sternocleidomastoid (12.0 ± 8.8 / 6.8 ± 7.8), diaphragm (11.7 ± 8.1 / 5.4 ± 7.1), and rectus abdominis (10.4 ± 4.7 / 3.1 ± 1.9) muscles. RMS attenuation was significant for all muscles (p < 0.001), with no agreement between pre- and post-filtering from Bland-Altman analysis. Conclusions: The applied digital filter attenuated RMS values while maintaining the same respiratory muscle activation, enhancing signal identification by reducing noise.