Dear Will, Tali and however is interested,
without having the paper nor a sharp memory of it at hand, I think if you
have an interleaved slice acquisition, an old-fashionedly long TR and
already performed spatial smoothing, you may still run into problems as you
mix voxels across time and space (it's a bit like Star Trek).
Somebody please let me know if that's not so.
Best wishes,
Helmut
A note/question
----- Original Message -----
From: "Will Penny" <[log in to unmask]>
Sent: Friday, April 04, 2008 3:37 PM
Subject: Re: Slice time correction is DCM - only in SPM5
> Tali Bitan wrote:
>> Dear DCM experts In the previous version of DCM (in spm2) it has been
>> shown to be insensitive to slight changes in timing and it was indicated
>> that DCM does not require slice timing correction. However, in SPM5,
>> slice-timing correction is an inherent stage of the DCM specification.
>>
>> My question is:
>> 1) Why is slice-timing correction needed for DCM in SPM5 (when it was not
>> necessary in SPM2)?
>>
>
> SPM2-DCM is robust to timing misspecifications of up to 1s. Given your
> central slice as a reference this means it is suitable for TRs<=2s.
>
> SPM5-DCM is given information about slice acquisition and can handle TRs
> up to 5s.
>
> Have a look at Stefan Kiebel's paper:
>
> doi:10.1016/j.neuroimage.2006.10.026
>
>> 2) What should we do during DCM specification if our data is already
>> slice-
>> time-corrected during pre-processing?
>
> If you have a TR < 3s I would do nothing, otherwise I think SPM5-DCM
> will probably be a lot better than the interpolation used in the
> 'slice-time correction' offered in preprocessing. Stefan has simulated
> just this comparison in his paper so its best to have a look at that.
>
> Best,
>
> Will.
>
>> Thanks a lot
>> Tali Bitan
>> Haifa University
>>
>>
>
> --
> William D. Penny
> Wellcome Trust Centre for Neuroimaging
> University College London
> 12 Queen Square
> London WC1N 3BG
>
> Tel: 020 7833 7475
> FAX: 020 7813 1420
> Email: [log in to unmask]
> URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
>
|