[FieldTrip] Source localization of MEG resting state data
tzvetan.popov at uni-konstanz.de
Thu Jun 11 07:58:48 CEST 2015
Hi MK CHOE,
your second attempt look reasonable at least to me. The topography suggest parietal activation and you have peak maxima over parietal regions.
> Hi, Tzvetan.
> Thank you for your first response.
> The weird output that you asked in first response is first attached figure 1.
> I thought It was weird because the sensor level output is so different form the source level output.
> I revised the script regarding your first advise and following tutorial. The output is second figure.
> It is better than first output. However, the sensor level output correspond with the source level output, exactly.
> Is it anything that need to revise?
> If the data is plotted at the source and sensor level, What the difference between two plots can emerge in general?
There isn’t an easy answer to this question. Sometimes you’d use to extract time courses out of your data without “seeing” any patterns on sensor level due to noise. Other times the sensor topography
almost perfectly matches the source map. That said sensor/source correspondence is good but if you lack this it doesn’t necessarily invalidates the results you just need to have plausible explanation of
why you don’t see things on sensor level while you believe they must be there. This is typically in cases of simultaneous recordings of various types- EEG/fMRI, tAC/DC-MEEG etc. Finally, while your sensor
signals are linear superposition of all source activity while your beamformer approach de-mixes those it is a “good sign” that sensor and source results are not mirroring each other.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the fieldtrip