Welch Allyn Means ECG Physicians Manual for CP Series Electrocardiographs - User Manual - Manual (Page 5)

MEANS Physicians Manual
Welch Allyn
5
1.2 Pattern recognition
This part deals with the analysis of the various waveforms. First of all, the QRS complexes
must be detected. No other waves or artifacts should be labeled as such. The intervals
between QRS complexes are measured and stored. After all QRS complexes have been
detected, they are typified, i.e., a comparison is performed that gives rise to classes of similar
QRS complexes.
Often there is only one type of QRS complex. If there are more, the “ordinary,”
“representative” or “dominant” one is established; the others are “extraordinary” or “non-
dominant”. Mostly, the number of dominant complexes in a recording is larger than that of the
non-dominant ones. In special cases this may not be true. In bigeminy their number may be
equal to that of the non-dominant complexes, or be one less or one more, depending on when
the recording starts and stops. If runs of tachycardia occur, the unusual complexes in a
recording may even outnumber the dominant ones.
The second step is to search for atrial activity. Both P waves and flutter waves can be
detected, when present. PP and PR intervals are also measured and stored for use in the
rhythm analysis.
The third step is to mutually compare the ST-T segments of the dominant complexes. For the
calculation of the averaged complex, only complexes are selected that have not only similar
QRS, but also similar ST-T. In this way complexes that are disturbed by spikes or sudden
baseline jumps are discarded.
For the morphological analysis, the selected dominant P-QRS-T complexes are averaged into
one complex. The main advantage of averaging is to improve the signal-to-noise ratio. Noise
is random and, in the averaging, the positive and negative oscillations will cancel out. An
additional advantage is that the analysis now has to be performed only once, i.e., on a single
representative complex. It may occur that in the averaged complex a P wave appears which
was not consistently detectable in the rhythm analysis, or vice versa.
The final step in the pattern recognition process is the determination of the zero level in the
representative P-QRS-T complex and the identification of the points of onset and offset of P,
QRS, and T. The zero level is determined for the averaged complex per lead in an interval
preceding the onset of the QRS complex. Onsets and offsets however are determined
simultaneously over all leads together.
1.3 Parameter extraction
After the onset and end points of P, QRS and T waves have been established, the relevant
parameters can be measured to provide the input for the diagnostic logic. Besides amplitudes
and durations, other measurements such as surface areas under the signal are derived. Most
measurements are made on the averaged complex in each lead separately (e.g., R
amplitude, Q duration), but some are derived taking all leads into account (e.g., overall QRS
duration, PR interval). These durations are generally longer than one would measure by hand
in individual leads or lead groups since the first onset in any lead and the last offset are taken
into account.