Automatic assessment of electromyogram quality . C., Sinderby, L. Lindstr[diaeresis]om, A. E. Grassino. Department of medicine, Notre Dame Hospital, University of Montreal and Meakins-Christie Laboratories, McGill University, Montreal, Quebec, Canada, and Spinal injuries unit, Sahlgrenska Hospital, G[diaeresis]oteborg, Sweden, and Department of medical information Processing, Sahlgrenska Hospital, G[diaeresis]oteborg, Sweden
APStracts 2:0286A, 1995.
Power spectrum analysis of the diaphragm electromyogram (EMGdi) is time consuming and no criteria have been developed to objectively quantify contamination of the signal. The present work describes a set of computer algorithms which automatically select EMGdi free of the electrocardiogram (ECG), and numerically quantify the common artifacts which affect the EMGdi. The algorithms were tested: 1) on human EMGdi (n=5) obtained with esophageal electrodes positioned at the level of the gastro-esophageal junction, 2) on EMGdi obtained in mongrel dogs (n=5) with intra-muscular electrodes in the costal diaphragm and 3) on computer simulated power spectrums. For both authentic and simulated power spectrums, indices were obtained by the algorithms, and were able to quantify signal disturbances induced by noise, electrode motion, esophageal peristalsis (in humans) and non -QRS complex related ECG activity. With the index incusion thresholds set to levels which allowed for a high signal acceptance rate with relatively small artifact induced fluctuations (10-15 %) of the EMGdi center frequency, the computer algorithms were found to be either equally or more reliable than other methods, including careful visual selection of the time domain signals by experienced analysts. In conclusion, the frequency domain application of computer algorithms offers a reliable and reproducible means to objectively quantify the sources which contaminate the interference pattern EMG.

Received 16 September 1994; accepted in final form 12 June 1995.
APS Manuscript Number A970-4.
Article publication pending Journal of Applied Physiology.
ISSN 1080-4757 Copyright 1995 The American Physiological Society.
Published in APStracts on 11 July 1995.