Dear Davy,
2009/12/8 闫伟 <[log in to unmask]>:
> Hello SPMers!
> 1, I want to know when doing MEG analysis, if I don't correct
> baseline, what will it have effect on the result?
If you apply high-pass filtering there is no need to do
baseline-correction as well. The point is that you shouldn't have huge
DC offsets in your data. The effect they might have depends on your
analysis but I don't see any analysis in SPM where leaving constant DC
shifts in the data would yield any benefit. I can, however, think
about a number of cases where these DC-shifts will cause problems.
> 2, For MEG data discribed in the manual of SPM8b, I hypothesis
> that there were 10 subjects, MEG data was detected while everyone
> viewing 86 faces and 86 scrambled faces. I want to do the 2nd
> level t-test for the two conditions(time 100-200ms, frequency 8-
> 25Hz). So I do 3D reconstruction and writed smoothed or unsmoothed
> *.nii files and used sw**.nii file to do the 2nd test, is there
> something wrong?
Sounds OK.
> 3, If I want to do 2nd test such as Condition_face (100_200ms
> subtract the baseline -100_0ms) compare to Condition_Scramble
> (100_200ms subtract the baseline -100_0ms), and produce 4 type
> sw**.nii file such as:
>
> sw_face_100_200.nii;
> sw_face_-100_0.nii;
> sw_scramble_100_200.nii;
> sw_scramble_-100_0.nii;
> How can desige the matrix?
>
> Whether I use Imcalc function (sw100_200.nii subtract sw-
> 100_0.nii)to get the image and use this image to finish the group
> analysis?
>
I think using ImCalc be the simplest option in this case.
Best,
Vladimir
|