Power Spectral Density - how can I get a link back to the original signal

[Dr Ingend]

Hi everyone!

I would like to analyze a biological signal for periodicities using a Fourier transform, i.e., I'd like to get a frequency distribution plot based on a long stretch of signal. Additionally, I would like to be able to mark the positions of certain frequencies within the signal that give rise to a peak in the freq distribut plot.

Do you have any suggestions how to perform such an analysis?

Thanks in advance!

 

[nasu]

If you did FFT of the whole signal, you cannot do this.
Each point of the FFT spectrum depends on all the points of the time domain signal.
You can do the so call "windowed" FFT when you cut your signal in several segments and do FFT transform for each. Or you use a mowing window and see how the spectrum evolves as you move the window.

 

[Dr Ingend]

Thank you! I thought about a moving window as well, but wasn't sure how to actually make it work. I think I would have to write it in Matlab from scratch.

Cheers

 

[reasonableman]

The Matlab signal processing toolbox will have the functions you need, but it doesn't take long to write them.

The 'official' name is a short time Fourier transform:
http://en.wikipedia.org/wiki/Short-time_Fourier_transform

 

[PJC01]

Hello,

Thank you for your question about analyzing a biological signal for periodicities using a Fourier transform. The power spectral density (PSD) is a useful tool for this type of analysis, as it provides a frequency distribution plot of the signal.

To get a link back to the original signal, you can use the inverse Fourier transform. This will allow you to convert the frequency domain representation back to the time domain, giving you the original signal with the marked positions of certain frequencies.

In terms of suggestions for performing this analysis, I would recommend using a software program or programming language that has built-in functions for calculating the PSD and performing the inverse Fourier transform. Some examples include MATLAB, Python, and R.

I hope this helps and good luck with your analysis!