Dear Michele
> first of all I swear this is the last question for you (at least for
this week).
Sure?
> Now, we succeeded in obtaining the correct behaviour for the z axis.
What we did
> is to maintain the down-up images (as registered from the nmr) both in
the
> functional and in the structural cases; then we performed the
realignement and
> smoothing steps; then we coregistered with the t1 image and everything
went
> through.
This is fine.
> Now the question is: in the statistical analysis we obtained that the
signal to
> noise ratio after coregistration with the t1 image is much greater
than the snr
> without coregistration. Could you please explain us whether this is
correct and
> why?
Do you used the structural image as the target and the functional images
as the objects in the coregistration step? In this case, the resliced
functional images will have the same voxel size like the structural
image, which have normally a higher resolution. So, the interpolation to
this new image resolution will result in a smoother image than the
original one. This may be the reason, for the difference between the two
statistical analysis. I would prefer the other way round. Use the
functional images as the target and the structural image as object and
choose the coregister only option, only a ***.mat-file will be written
to the structural image. The overlay procedure will use this mat-file,
so you do not lose any image resolution in the structural image.
Is this the answer to your question?
Good luck and a nice weekend.
I'm looking forward to new questions next week...
Karsten
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|