I have had some queries off-mailbase about whether I got a response to
this and attach a reply sent by Rik Henson (I hope this is OK Rik):
Mike -
In the face example data, reslicing is necessary if you intend to
perform subsequent slice-timing correction (because nobody can be
bothered to rewrite slicetiming correction to "interpolate" on the
fly!).
I don't think it is necessary for the auditory dataset, other than to
write a mean image: you could ask Will to change this, because he wrote
that chapter, but I doubt it is that important.
As far as I am aware, the same general advice of avoiding reslicing as
much as possible applies to SPM5 the same as SPM2.
Rik
----------------------------------------
Dr Richard Henson
MRC Cognition & Brain Sciences Unit
15 Chaucer Road
Cambridge
CB2 7EF, UK
Tel: +44 (0)1223 355 294 x522
Fax: +44 (0)1223 359 062
http://www.mrc-cbu.cam.ac.uk/~rik.henson
----------------------------------------
>-----Original Message-----
>From: SPM (Statistical Parametric Mapping)
>[mailto:[log in to unmask]] On Behalf Of Mike Glabus
>Sent: Thursday, March 01, 2007 1:39 PM
>To: [log in to unmask]
>Subject: Re: [SPM] realign and normalize with SPM5
>
>I just want to pick up on this as there is another fundamental
>difference
>in preprocessing as outlined in the SPM5 manual, compared to SPM2.
>
>From what I understand, it was recommended by the authours
>that for SPM2,
>the fMRI time series is realigned only and not resliced: the
>realignment
>parameters being saved in the .mat file for use in the next
>stage, spatial
>normalisation, (effecting resampling and reslicing in a single step).
>
>The recommnded preprocessing procedure for SPM5 (according to
>the example
>data sets in section VI of the manual) is to realign *and* reslice the
>fMRI data prior to spatial normalisation. Is there a resaon
>for adopting
>this approach in SPM5? Would it be acceptable to simply estimate the
>realignment parameters, as they would (presumably) be
>considered during
>the spatial normalisation procedure? This would save a lot
>disk space used
>for writing out rr*.img files.
>
>Regards - Mike
>
|