I have two EEG signals that are very similar. The difference is only in amplitudes. However, they are coming from two different cognitive processes. What are some methods, beside FFT, for distinguishing between two very similar signals in terms of statistical information? Please let me know. Thank you


  • 1
    $\begingroup$ When you say that difference is only in amplitudes, what exactly do you mean? Do they have the similar spectral content (alpha waves, delta waves, etc.)? Perhaps you can post some images or data that you're analyzing. This will help us see how exactly the signals are different. $\endgroup$
    – Phonon
    Feb 20, 2012 at 3:34
  • 3
    $\begingroup$ @BrainStorm Funnily enough your sign off alias 'BSS' is also the acronym for a methodology that might help you - BSS - also knows as 'Blind Source Separation', also somewhat related to deconvolution. Ill echo Phonon's request for more info, but sans that, you can also look into 'ICA' (Independent Component Analysis). ICA will basically separate a number of signals that have been 'mixed' (linearly combined) together, based on the assumption that they came from independent processes, (which if they are made by different parts of the brain might fit that model). $\endgroup$
    – Spacey
    Feb 20, 2012 at 5:48

2 Answers 2


yet another thing to try could be wavelet transform. Wavelets are powerful tools for analysis of signal. You could take a DWT of your signal and compare may be the most significant coefficients. Anyway but it would be helpful if you could post some sample data.


I'll be surprised if the difference is only amplitudes. And if they are, just take the mean amplitude of both signals and you know which one is which. However, I doubt if thats the case.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.