Dear FSL Experts,
I am having difficulty running group-melodic (Multi-session temporal
concatenation). Particularly, a group.ica design file that I created
consistently crashes when searching for a hdr file (bg_image) within an
individual participant's group.ica directory. The specific error reads as
follows:
** ERROR (nifti_image_read): failed to find header file for
'/mnt/plum3_share3/2009/09-05-12.1/fmri/Resting/physio_corrected/smoothed_spm5_filtered_high/swar_group.ica/reg_standard/bg_image'
** ERROR:
nifti_image_open(/mnt/plum3_share3/2009/09-05-12.1/fmri/Resting/physio_corrected/smoothed_spm5_filtered_high/swar_group.ica/reg_standard/bg_image):
bad header info
Error: failed to open file
/mnt/plum3_share3/2009/09-05-12.1/fmri/Resting/physio_corrected/smoothed_spm5_filtered_high/swar_group.ica/reg_standard/bg_image
Cannot open volume
/mnt/plum3_share3/2009/09-05-12.1/fmri/Resting/physio_corrected/smoothed_spm5_filtered_high/swar_group.ica/reg_standard/bg_image
for reading!
This error occurs in Higher-level MELODIC, after the Feat main script has
run for many hours. The strange thing about this error is that the
individual participant’s group.ica directory itself was created by the Feat
main script, because of which I assumed that the hdr file (bg_image) that
the script is looking for within the group.ica directory should have been
created by the Feat main script as well.
The data we are using for this group.ica analysis was preprocessed in SPM
and compiled into 4D niftis. Because the data has already been smoothed and
normalized, the ‘standard space’ template we specified for registration in
group.ica is the same template we used when preprocessing the data in SPM
(to avoid renormalization). Because our data is already in standard space,
is registration still necessary (if not, is there some way to deselect the
‘standard space’ option? For some reason, the GUI is resistant to this)?
Also, is it possible that either (1) the data we are using for group.ica is
incompatible with FSL simply because it was preprocessed in SPM, or (2) the
standard template we used in SPM is incompatible with FSL? And finally, is
it possible to relaunch this design file with only the Higher-level Melodic
and not the Feat main script?
Thank you for helping me diagnose this problem that I am experiencing.
Please let me know if any further information will be helpful.
Best,
Lucina
|