Dear Vladimir,

I am analyzing resting state EEG data in scalp-frequency space. I have had success with this procedure:

1.) On epoched & preprocessed data, I computed the Hilbert transform.

2.) when running "convert to images", I averaged over the frequency band of interest.

3.) then I entered these scans into the 2nd level design.

However, is it possible to use FFT instead of a time-frequency estimation method?

Since we are not utilizing the time dimension in the eventual SPM, can we drop time with an FFT and reduce the data to 3D (two spatial dimensions and FFT power)? 

This is not as much for a practical reason (besides processing time) as to reduce complexity that must be explained in a review.

Thanks for your time.

Best,

Brian

Brian Erickson

Ph.D. Candidate, ACBS Program

Drexel University Department of Psychology

College of Arts and Sciences