> we discovered serious image acquisition errors in
> 3 volumes out of a time series of 240 scans.
> It must be due to a mistake of our scanner,
> these 3 volumes are definitely not usable,
> because they only contain white patterns or stripes.
> Because after closer inspection we discovered
> artefacts like this in 4 subjects, we would not like to throw away these
> data.
>
> What is the best way to deal with things like that ?
>
> We could delete the single volumes and just continue
> like if they were never acquired, but this would not match
> the assumptions about the timecourse of hemodynamic responses.
>
> Another way would be to replace the volumes by the mean
> of the volumes acquired before and after the affected one.
>
> Any other ideas ?
It's not really my area and I may be wrong, but one possibility is to
include the dodgy scans in the analysis but add an extra column to the
design matrix for each of them. Each column should contain all zeros,
except for a one correspondingto the bad scan. These columns should not
be convolved with the HRF. I don't know which particular SPM fMRI design
this would correspond to though.
Best regards,
-John
--
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
|