PSD window length, and freq range

[ariyas]

I have a 15-minute continuous resting-state MEG recording imported into the database. For source-level PSD analysis, is a 66-second window appropriate for examining frequencies as low as 0.03 Hz, considering it would only capture about two cycles (or would it be 4 cycles .. with a 50% overlap)?
A 33-second window length would be too short?

Also, since the MEG acquisition bandwidth was 0.03–330 Hz, any PSD content below 0.03 Hz (from longer windows) and above 330 Hz should be noise/artifacts/drifts. Is there a way to hide that or not have those values when exporting a Scout PSD to MATLAB (when using MATLAB's FFT defaults and not grouping in frequency bins)

Thanks

[Raymundo.Cassani]

@ariyas

A 33-second window length would be too short?

Yes, that is quite short, 33-second window implies only one cycle at 0.03Hz.

@ariyas

Also, since the MEG acquisition bandwidth was 0.03–330 Hz, any PSD content below 0.03 Hz (from longer windows) and above 330 Hz should be noise/artifacts/drifts. Is there a way to hide that or not have those values when exporting a Scout PSD to MATLAB (when using MATLAB's FFT defaults and not grouping in frequency bins)

Yes, there is a way. Once you have computed the PSD, on that file run the process Extract > Extract values and indicate the desired frequency range. This will create a new PSD file defined for that range.