Thaks a lot for the answer
>I guess this is where all the problems start. SPM can't find
>any interesting data to analyse.
>SPM loops over voxels in a slice by slice fashion.
>See line 592 (of spm_spm) (for z=1:zdim) - are the dimensions of your >images
>specified correctly ? (if not SPM may not go thru all the slices and
>may not find any inmask voxels).
>SPM does'nt bother to anaylse voxels where values don't
>vary over time (see lines after comment '% mask our voxels where data is
>constant'). You could use
>'display' to check your images show some variation.
>Also, if you specify a mask yourself (SPM calls this an explicit mask) SPM
>only looks at these voxels. But it seems you have'nt specified a mask.
>Is this correct ?.
It's correct
We don't understand the way voxel are implicitly thresholded in spm_spm.m
line 616
Cm(Cm) = Y(i,Cm) > xM.TH(i);
what's the meaning of xm.TH matrix and when is it set?
we have void mask because ex. mean(Y(1,Cm)) ~= 1.9 and xM.TH(1) = 80
what's the meaning of this low values of Y?
And also ,plotted time series show a pattern ,so they vary over time...
Thanks for any advice
Marco
|