Identifying QRS point in ECG signal

Hi guys, I'm trying to find the QRS point from the ECG signal that I had been given. I had found the R peak, however, I couldn't code the Q and S point. Can someone please help me to check my coding if there is something wrong with it?

load('Sample_1.mat')

%The recording were digitized at 360 seconds per second with 11 bits
%resolution over mV range.
length(Orig_Sig) = 3600;
D = 1:3600;
t = D./360;

Fs = 360;

%plot noisy signal 
figure
subplot(211), plot(t,Orig_Sig);
title('Original ECG Signal')

%lowpass Butterworth filter
fNorm = 25 / (Fs/2);
[b,a] = butter(10, fNorm, 'low');
y =filtfilt(b, a, Orig_Sig);
subplot(212), plot(t,y);
title('Filtered ECG Signal')

[R1,TR1] = findpeaks(y,t,'MinPeakHeight',1000);
[Q1,TQS1] = findpeaks(-y,t,'MinPeakHeight',-850,'MinPeakDistance',0.5);

figure
plot(t,y)
hold on
plot(TR1,R1,'^r')
plot(TQS1(1:2:end), -Q1(1:2:end), 'vg')
plot(TQS1(2:2:end), -Q1(2:2:end), 'vb')
hold off

[Huy Le Van]

Hi. Follow me, you can be reference this code.

Answer 1

[Star Strider]

That is an interesting EKG. It displays frequent unifocal PVCs with obvious retrograde conduction, blocking the subsequent P-wave.

As much as I like findpeaks, the islocalmin function might be more appropriate, since you can use the ProminenceWindow name-value pair, and others not available in findpeaks. That might be easier with respect to locating the Q and S waves. In any event, the findpeaks MinPeakProminence parameter with or in place of MinPeakHeight may be more valuable than MinPeakHeight alone.

Also, the 25 Hz lowpass filter cutoff frequency is a bit restrictive and could obscure some detail. Experiment with a 45 Hz cutoff and see if that improves your ability to detect the Q and S deflections.

[Stella Sim Jun Ni]

[Q1,TQS1] = islocalmin(y,'MinProminence',80);
plot(t(Q1),y(Q1),'vg');

I tried to code it this way, which it did determined the Q wave/S wave.

The problem now is that 2 of the S wave is not on the right point. And also, I'm stuck on not being able to plot both Q and S point using different point colour, like before using the function (1:2:end).

I also have to find both P and T point.

It'll be great if you can give me some guidelines. I'm new to this programming stuff.

Thanks a lot.

[Star Strider]

Since the Q-waves are always the first, and the S-waves are always the second, from the image you posted, first use the find function to create numeric vectors from the logical indices, then define them as:

Q1idx = find(Q1);                       % Numerical Indices
Qdef = Q1idx(1:2:end);                  % Q-Deflection
Sdef = Q1idx(2:2:end);                  % S-Deflection

You may need to reverse those, since I cannot test this to be certain that is correct, however from the image you posted, that should work. Plot them separately so they are different colours, the same way you did the current plot, only using those indices instead of ‘Q1’.

[Stella Sim Jun Ni]

It works perfectly fine. Thank you for the help! Really appreciate it :)

[Star Strider]

As always, my pleasure!

[Ilmiat]

Hi

can you kindly tell me how to find P wave and T wave after that? on this same signal?

thanks in advance