filter Spikes from lung recorded signal

Hi , im a last year student . for my final project , im trying to record and analyze lung sound. i have successfully record the sound , but it has spikes and another additional noise . my question is, how can i filter these spikes, using filter function of matlab.

[Star Strider]

‘Spikes’ are high-frequency, broadband noise, but then so are most lung sounds. Your first step is to do an fft, or preferably a spectrogram, of your signal so you can visualise the frequency, or preferably time-frequency, components then design your filter to pass only those frequencies of interest. The ‘spikes’ you circled could be end-expiratory wheezes, which are important in diagnosing asthma and chronic obstructive pulmonary disease (COPD). (This is when the small airways close at the end of the expiratory phase of the respiratory cycle. They should stay open, so the wheezes are indicative of a loss of lung elasticity in COPD, or the presence of small airway inflammation in asthma.) Not everything that seems to be noise in biomedical signals is noise. Some of it is important information.

[cob]

these spikes which i mentioned are executed when the diaphragm touch and leave the skin. i'm sure that these are not part of a signal which i need.

thank you in advance .

Answer 1

[Star Strider]

‘these spikes which i mentioned are executed when the diaphragm touch and leave the skin. i'm sure that these are not part of a signal which i need.’

I’m certain that you mean stethoscope diaphragm. That would have been helpful to know in advance.

My advise remains the same: first analyse your signal with a spectrogram to see the time-frequency content of the signal. Stethoscope diaphragm contact noise likely has a different frequency content from lung sounds, but you won’t know the frequency band until you do the spectrogram. That will let you identify the specific frequencies of the diaphragm contact noise because you can isolate it by time as well. Then do the fft to see the overall spectrum. You will then know if you can filter out most of the unwanted signal.