Does it make sense to do any kind of time-frequency analysis on averaged ERFs?

Suppose you have a simple time series, averaged from a multitude of trails—does doing an FFT or other methods have any purpose?

It depends on what you are looking for in the data. A lot of information in EEG is contained in the single-trial dynamics and therefore is lost in the time-domain averaging. So if you are starting with the time-domain average (that is, the ERP), then I’m not sure what the FFT would tell you that isn’t already in the time domain.

You could certainly try it – it’s a perfectly valid approach – but the question is whether it will give you more information. A better approach would be to apply the FFT (or time-frequency analysis) to each individual trial and then average the spectral responses together.

1 Like

I suppose that’s what I was wondering—if it gives any more information. I agree, I think time-frequency analysis would be nice, since I’m working with a paradigm looking at the response time of an auditory stimulus.

Thanks for the input!