i have an ecg signal image, how to i get the parameters of the same?

3 views (last 30 days)
this is my image wherein i need to know the duration at qrs wave.
<<
>>

Accepted Answer

Star Strider
Star Strider on 18 Apr 2015
First, the time base is in seconds, not milliseconds. You can verify that by measuring the single R-R interval (0.85 s to my calphers).
The record is a Lead II rhythm strip. Normal sinus rhythm, rate 71 bpm. Both complete QRS complexes are 0.2 seconds (upper limit of the normal QRS is 0.12 seconds) with prominent S-waves. The QT segment is 0.38 seconds, approaching the upper limit of normal (0.4 s). The PR interval is 0.13 (normal 0.2 s). The T amplitude appears to be high. No pathological Q-waves or ST-T segment abnormalities are seen. The U-wave is prominent.
IMPRESSION: Abnormal EKG with short PR interval, intraventricular conduction delay characterised by pathologically prolonged QRS complexes, normal QT, elevated T-wave, and prominent U-wave consistent with hyperkalaemia, hypocalcaemia or drug effect. Suggest clinical correlation.
— Star Strider, M.D., Diplomate, American Board of Internal Medicine
  2 Comments
SANJANA GANESH
SANJANA GANESH on 18 Apr 2015
how do i get the duration of any ECG signal? any specific code can be written?
suppose i scan two ECG's of the same person at different heart rates, how to prove that both are of the same person?
Star Strider
Star Strider on 18 Apr 2015
From that particular EKG, (I don’t have anything other than the trace you provided), you would need to detect the peak (easy, you can threshold it), then trace it down both sides until you encounter a flat isoelectric region. The change from the flat isoelectric region to the beginning deflection of the R-wave is the onset of the R-wave, and the beginning of the first isoelectric segment after the S-wave is the end of the S-wave. Those transitions mark the beginning and end (respectively) of the QRS complex. (Note that isoelectric means a flat region at 0 Volts.) You might be able to use the gradient function to help you with that. Other EKG characteristics (P, T, U) have been studied by those familiar with wavelets, which seem to be the most reliable way to detect them. ST-T wave elevation or depression are even more difficult to detect.
You can write specific code, however, EKG traces differ by patient position (because the heart changes position relative to the surface electrodes) so even in the same healthy patient at rest, the EKG can change. Before you start your project, search PubMed to see what others have done. (This is not a new problem, and some of the techniques are quite sophisticated. Start with the more recent review articles — especially the IEEE articles — and work back to the original research articles.)
I doubt that you could ever prove that two EKGs are from the same person unless they had specific identifying characteristics (old myocardial infarct, conduction disorder, or other abnormality) and only one person in the population had them in that particular configuration. It’s generally impossible even to tell the gender of the patient from an EKG, much less any other identifying information. Different heart rates would make it more — not less — difficult to determine if the EKGs were from the same person. Besides, the neurohormonal influences that change the heart rate change other aspects of the EKG.
Nothing in medicine is ever straightforward, and the myriad biomedical engineering approaches to any problem reflect that.

Sign in to comment.

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!