Hi,
I performed a power spectrum analysis on an epi scan with 84 repititions using the afni power spectrum plug-in. The attachments "time_series.pdf" and "spectrum.pdf" shows the intensity versus time and the power spectrum, respectively, of one voxel. That power spectrum cannot be real. The real power spectrum of the time series should have a very sharp maximum at f = 0, which I verified using Matlab.
Any idea what might be the reason? Has anyone encountered similar problem before?
Thanks,
Xiaopeng