Hi,
As Steve said, there are also possible extension fields in the header, which can be of very general format and so can possibly contain sensitive information. These are not shown by fslhd. However, you can remove them by running the file through fslmaths: e.g. fslmaths t1_con_001 t1_con_001 (obviously you can use any name, but if you use the same name for input and output then it will simply run it through the FSL input/output library which will strip extensions). Doing this will also replace the description field (one place where such information could also appear) with "FSL5.0". At that point it is really only the aux_file field (which can be checked with fslhd), the filename itself and the face in the reconstructed image that could contain any identifiable information.
All the best,
Mark
On 15 Oct 2014, at 08:43, Martin Dyrba <[log in to unmask]> wrote:
> Hi,
> To extend the previous message a bit:
> In contrast to DICOM, the NIFTI standard (http://nifti.nimh.nih.gov/) does not contain any personal information about the subject. That means, when converting DICOM to NIFTI, all those information get lost (except maybe the subject name that will be used as file name for the NIFTI file). To make sure, check the header field note/description and DB name, which might be used by some programs to store additional information.
>
> If you are interested in defacing the scans, I can recommend mri_deface (https://surfer.nmr.mgh.harvard.edu/fswiki/mri_deface , also part of FreeSurfer) for T1 weighted images.
>
> Regards,
> Martin
|