Enter the fMRI as different sessions into the realignment.  This will
deal with stuff automatically.  Each scan in a session will be registered to
the first scan of the session, and all sessions will be aligned with each other
by matching the first scan of each session to the first scan of the first session.
As for registering the structural and functional images, I would coregister the
structural images together, and then match the structurals to the functional
images.  I would suggest using the Mutual Information method for doing this
(get the latest patches), selecting a functional as target, a structural as
source, and the other structurals as source.  Ideally the source and target
images would be mean images, but this isn't strictly necessary.

Best regards,
-John

> we are using normalized epi-sessions as functionals and take a stuctural
> image on each aquisition day (so we can use the positional information as
> starting point for coregistration and use the structurlal image to get the
> normalization parameters).
>
> If I want to average severall data sets of the same subject over severall
> aquisition days I can:
>
> 1) corregister the functional data of day2 (and other days) to the
> structural scan of day one and use the day1 normalization parameters.
> 2) corregister the functional data of day2 with structural data of day2.
> Then corregister structural data from day2 with structural data from day1
> (and the functionals along with that), then use normalization parameters of
> day1 3) normalize data (functional and structural) of both days separately.
> 4) ... many more possibilities
>
> Though in theory equivalent there might be practical experiences which is
> (at least with SPM99) the best way to realign all normalized functional
> data over severall aquisition days.
>
> Any tips which is the best way (or which way doesn't work reliably)

--
Dr John Ashburner.
Functional Imaging Lab., 12 Queen Square, London WC1N 3BG, UK.
tel: +44 (0)20 78337491  or  +44 (0)20 78373611 x4381
fax: +44 (0)20 78131420  http://www.fil.ion.ucl.ac.uk/~john