Dear Marc and Beatriz,
>Related to this issue, we have a subject who moved between the anatomical images and the first and second fMRI runs. When we reorient, realign, and normalize EPI images to the T1-anatomical images, the result is obviously wrong, and we get a shifted map of activation, where areas are not correctly located.
>We have tried to solve this problem by reorienting, realigning, and just normalizing EPI images to themselves. However, we seem to get a rather similar, unsatisfactory result.
>Is there any better way to deal with this?
>
>
Yes, I think so.
First with regards to Marc's questions. I will assume that by runs you
mean sessions. SPM will start by taking the first volume from each
series, and align them to the first volume of the first series. After
that it will align scans 2:n of each series to the REALIGNED first scan
of that series. Hence, after realignment all scans from all sessions are
(hopefully) in the same space.
For Beatriz; I think that you should
1. Realign all scans (for all sessions) for a given subject chosing
either "Create all images + mean" or just "Create mean", depending on if
you want to use Unwarp/Slice timing.
2. Do a "Coregistration" chosing your mean EPI image as "target" and
your anatomical scan as "object". You may or may not want to reslice at
this point depending on if you plan to proceed with "Spatial normalisation".
Good luck Jesper
|