Dear Muhammad,
On 27 Feb 2010, at 22:04, MP wrote:
> Hello SPMers,
>
> I was wondering if there is an explicit way to calculate ERS and ERD
> of certain EEG frequency band in SPM? I know that ERS/ERD can be
> calculated as the percentage
> of a band power change post-stimulus onset with respect to the pre-
> stimulus baseline or a ‘resting’ interval. For this should I carry
> out spectral analysis separately for pre- and post-stimulus epochs
> and then create a contrast between the two at the 1st level?
>
You can use Other/Rescale TF (spm_eeg_tf_rescale) with either Rel (to
get % baseline) or LogR (to get dB) option.
> Secondly, if i am interested in doing a spectral analysis of my EEG
> data to see the differences in certain delta, theta and alpha
> frequency bands between two conditions, should I carry out one
> analysis with frequency range encompassing these three frequency
> bands, or should I extract the data for each frequency band
> separately (do time-frequency transformation within the frequency
> window of each band)?
If you have a specific hypothesis, you inference will be more
sensitive if you only test things which are relevant for this
hypothesis. You can make time-frequency matrices for all the bands but
then use the 'small volume correction' option to only test the parts
you are interested in. The important thing is you should decide about
the things you want to focus on in advance and not based on the data.
Best,
Vladimir
>
> Any insight or any direction towards a better alternative will be
> much appreciated.
>
> Thanks
> -Muhammad
|