Signal Quality Indices for Ambulatory Electrocardiograms Used in Myocardial Ischemia Monitoring

Abstract

Ambulatory electrocardiograms (ECG) can be used to monitor patients for myocardial ischemia. Low ECG signal quality, due to contaminants such as motion artifact, can lead to an increase in false alarms leading to alarm fatigue. The false alarms can be reduced by processing only ECGs of ade-quate quality, quantified by a signal quality index (SQI); con-taminated ECG may be discarded. Four SQIs based on an esti-mate of the signal-to-noise ratio (SNR) were examined, where the mean, median, 25th percentile and minimum SNR were con-sidered. The SQIs were validated by contaminating 30 second segments of the record ‘s20031’ from the Physionet’s Long-Term ST Database with 30 second segments of the record ‘em’ from the Physionet’s Noise Stress Test Database. The SQIs of each segment were compared to the calibrated SNR of the segment. It was found that a strong correlation (Pearson correla-tion coefficient > 0.85) was exhibited between the SQIs and the calibrated SNR for each segment with SQI based on 25th percentile exhibiting highest correlation.Keywordsbiosignal quality analysiselectrocardiographymyocardial ischemiasignal processingsignal-to-noise ratio