Dear Wen-Ching,
> Is it true that the pre-processing procedures such as alignment,
> coregistration,
> smooth, slice timing, and normalization do not have to follow some kind
> of order to apply
> to fMRI data before statistical processing?
> Logically, there may be some kind of order, but mathematically, these
> procedures can be applied
> at any sequence order, and the results would not be "too much"
> different.
I'm afraid the order is not only about doing this in a logical order. In the
view you give above, all the preprocessing steps do something to the data, no
matter what you do first. This is not the case. After normalisation, a
coregistration will have no effect. After coregistration, alignment will have no
effect. All the operations above make use of the image being smooth, in order
for the optimisation routines to converge. I believe, most of the routines (when
operated from the GUI) take care of that themselves. So, unless you really know
what you're doing, stick to the normal procedure.
You are right in saying the results won't differ too much in the sense that if,
for example, a coregistration and a normalisation are performed, the result will
be a normalised image, regardless of the order in which you call them
(normalisation incorporates coregistration). But sometimes it is also necessary
to use one preprocessing step first, in order to guarantee another to be
successful. Aligned images normalise easier than raw images, so to say.
But most important is what you want to do with your images. The goal you
describe in your mail is to go from raw images to normalised images via any
sequence of preprocessing operations. I think you can do that by (smoothing and)
normalisation at once. In other cases, you might want to do your statistical
analysis on realigned or coregistered images.
This is what I understand of the preprocessing of images in SPM. Maybe you'll
some (more useful) comments from other people as well. I hope this explains it a
bit.
Best regards,
Alle Meije
--
Alle Meije Wink
Institute for Mathematics and Computing Science, room 105
University of Groningen P.O. Box 800
9700 AV Groningen
Telephone +31 50 363 71 27 The Netherlands
Fax: +31 20 875 48 00
E-mail: [log in to unmask]
WWW: http://www.cs.rug.nl/~wink
|