design a filter for heart signal

Hello I have piezoelectric heart signal and I want to separate the breathing signal and nonbreathing signal in MATLAB smartly(it means whatever signal I give it to that code it separates automatically), I'm thinking about a filter which can detect the high frequency and low frequency but I don't know how to do it. any suggestion?thank you.

[Jan]

This question does not concern the implementation in Matlab, but the scientific question how the signals can be distinguished. Better ask this in a forum about EMGs or search in the current literature.

[Christoph F.]

Is this a copy of

https://de.mathworks.com/matlabcentral/answers/360733-smart-code-for-separating-breathing-heart-signal-and-non-breathing-heart-signal

Technically, a regular filter will not work well. Pulse (heartbeat) signals start at about 0.67 Hz (40 bpm) or even 0.5 Hz if you want to measure all the way down to 30 bpm. (Relevant) respiration frequencies range from 0.05 Hz to about 3 Hz. There is a significant overlap between the two frequency ranges.

Of course, if you can exclude high respiration rates and low pulse rates, e.g. you operate under the premise of "pulse rate > 0.67 Hz" and "respiration rate < 0.33 Hz", using a simple lowpass filter might work to separate the two components.

It all depends on your application and requirements. You will need to research digital signal processing/digital filter design and application.

[shina mokhtari]

actually I have a signal data with time and I want to separate heart signal with breathing signal with matlab. I attached text file for timing and signal data(which is in column 2).

[Christoph F.]

A Butterworth filter with a corner frequency of 0.5 Hz and sufficient order should work for this example. The signal processing toolbox is very helpful here, since it contains filter design, analysis and application functions.

[shina mokhtari]

I don't know how to work with signal processing toolbox and I need it urgent, is there any way to learn quickly?

[Christoph F.]

Digital signal processing is a complex and extensive topic. Even with mathematical background knowledge, weeks or months of study are necessary.

You could start out with calculating a moving average of the signal with an averaging period between 640ms and 1000ms. This should show the low-frequency respiration component, but this method is in no way "smart".

If the task is urgent and important, paying a consultant to do it may be the best course of action.