Send to

Choose Destination
See comment in PubMed Commons below
Med Biol Eng Comput. 2000 Nov;38(6):659-65.

Identification of post-myocardial infarction patients with ventricular tachycardia by time-domain intra-QRS analysis of signal-averaged electrocardiogram and magnetocardiogram.

Author information

  • 1Helsinki University of Technology, Laboratory of Biomedical Engineering, Finland.


A new time-domain analysis method, which quantifies ECG/MCG intra-QRS fragmentation, is applied to parts of the QRS complex to identify post-myocardial infarction patients with ventricular tachycardia. Three leads of signal-averaged electrocardiograms and nine leads of magnetocardiograms were band-pass filtered (74 Hz to 180 Hz). The filtered signals showed fragmentation in the QRS region, which was quantified by the number of peaks M and a score S, that is the product of M and the sum of the peak amplitudes. Both parameters were determined for the first 80 ms of the QRS complex and the total QRS complex in each channel. For classification, the mean-values of the parameters M and S of the three electrical leads and the nine magnetic leads were calculated. Late potential and late field analyses were performed for the same signals. 31 myocardial infarction patients were included, 20 of them with a history of documented ventricular tachycardia (VT). Identification of VT patients using the SAECG led to better results (sensitivity 95%, specificity 91%) considering the entire QRS complex than with the standard late potential analysis suggested by Simson (sensitivity 90%, specificity 73%). For the SAMCG and the entire QRS complex results using the parameters S and M are also better (sensitivity 95%, specificity 100%) than for the late field analysis (sensitivity 90% and specificity 100%). For the first 80 ms, the performance of the parameters M and S is only slightly decreased.

[PubMed - indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Loading ...
    Write to the Help Desk