Hi Ada,
well I think your preprocessing is fine, just a few comments
> The first approach:
> Perform "slice timing" --> then perform "realign" (i.e. realign all
> runs, say 4 when there are 4 runs) and select default
> option "co-register + reslice" and then create "mean image only" -->
> after that, do "co-register" with the subject's anatomical image as
> the target and the mean image (created during the realign) as the
> source image, and then select all the 4 set of epi images (4 runs) -->
> after that, do "normalization" and use T1.mnc as the template image
> and the subject's anatomical image as the source image, and then write
> on all the co-registered epi images (4 runs) --> finally, perform
> "smoothing".
> The second approach:
> Perform "realign" and select default option "co-register + reslice"
> and then create "all images + mean images" --> after that, do
> "normalization" and use EPI.mnc as the template image and the mean
> image (created during the realign) as the source image, and then write
> on all the realigned epi images (4 runs) --> finally, perform "smoothing"
- Slice timing isn't always mandatory though I always do it (see the
list for discussion on this topic, specially for TR >~ 3s)
- The 'usual' way to pre-process is slice timing then realign for
interleaved acquisition or realign and slice timing or
ascending/descending orders
- co-register + reslice" and then create "mean image only" OK
- "co-register" with the subject's anatomical image as the target and
the mean image (created during the realign) as the source image, and
then select all the 4 set of epi images (4 runs) - this step is
optional, I'd rather take the EPI as target and T1 as source (and not
touching the EPI here) ; then normalize on the T1 template with my
coregistered subjectT1 as source and apply on the EPI
- "normalization" and use EPI.mnc as the template image and the mean
image (created during the realign) as the source image, and then write
on all the realigned epi images (4 runs) --> finally, perform "smoothing" OK
Some time ago I wrote a user guide for spm2 with the different steps and
the choice of options, etc ...
http://membres.lycos.fr/wamcyril/
hope this helps
cyril
|