AUTOMATED MEASUREMENT OF THE CYCLIC ACTIVITY OF RESPIRATORY MUSCLES

A microcomputer-based development for automated analysis of the electromyogram (EMG) recorded from respiratory muscles in a variety of situations is described. In addition, an assessment of the performance of the system is presented, along with data relating intercostal EMG activity to ribcage movement in seated subjects. Studies were performed on sixteen normal subjects - non smokers, mean (+/- SD) age 31 (+/- 6) years, mean (+/- SD) mass 78 (+/- 8) kg - of which fifteen studies proved suitable for analysis. Each study lasted for a period of five minutes, during which time recordings of intercostal EMG, ribcage postero-anterior displacement (RCPA) and airflow were made. For every breath taken by each of the subjects, the peak integrated EMG activity (iEMG(peak)) was measured both by hand and by the automated system. The automated and manual measurements of iEMG(peak), which ranged from 0.0 to 91.3 mu V, differed by only -0.82 +/- 3.34 mu V (mean +/- SD). The index iEMG(peak) and two additional indices of iEMG activity (iEMG(mean), iEMG(area)) were evaluated with respect to RCPA, a measure of overall respiratory activity. The indices of iEMG were observed to show an exponential dependence on RCPA(amp), the amplitude of ribcage motion. Following a log transformation to linearise the relationship, the correlation of each index with RCPA(amp) was evaluated; iEMG(peak) and iEMG(area) correlated similarly with RCPA(amp) (no significant difference at 5% level), but iEMG(mean) was found to be a significantly different (p < 0.001) and poorer correlate. We conclude that the automated analysis of respiratory iEMG as described in this paper can provide results showing consistency with manual measurement. The automated approach shows considerable benefit when substantial volumes of data are involved, or when it is desirable to perform analyses which are not amenable to manual methods.