frequency analysis of raw MEG

Hello!

I am a MNE/MEG beginner and I am interested in time/frequency analysis of
the raw (non-averaged) MEG data.

Can one deal with the raw data MNE output in the same way as with the raw
signals from the MEG sensors (e.g. to look at FFT spectrum and calculate
coherence between ROI)? Intuitively, looking at the time course of the
unaveraged MEG signal from a part of the source space (ROI) one should see
the physiological rhythms of the brain (e.g alpha rhythm). However, when
the unaveraged (ongoing) MEG is viewed in the mne_analyze, the time
courses at the separate vertexes do not resemble that at the sample
channel. The signals from vertexes have more of high frequency activity
and the waves are sharper than at the sensor signal.
Can somebody comment on that?

Elena