questions > Warning unknown Siemens slice order 41340006
Showing 1-4 of 4 posts
Jul 4, 2016 09:07 AM | Loreen Mamerow
Warning unknown Siemens slice order 41340006
Hello colleagues,
I got a warning (attached) during the dcm2nii routine (output to FSL/SPM8 4D NifTI nii format). Presumably this is ok and I can ignore it (?) but I wanted to make sure that all is well with my output files.
We scan on Siemens Prisma 3T, perhaps this is a source for the warning? The slice order gets recognized correctly.
Thank you in advance!
Regards, Loreen
I got a warning (attached) during the dcm2nii routine (output to FSL/SPM8 4D NifTI nii format). Presumably this is ok and I can ignore it (?) but I wanted to make sure that all is well with my output files.
We scan on Siemens Prisma 3T, perhaps this is a source for the warning? The slice order gets recognized correctly.
Thank you in advance!
Regards, Loreen
Jul 5, 2016 12:07 PM | Chris Rorden
RE: Warning unknown Siemens slice order 41340006
I am afraid this is a problem with your images and not my software,
but it is usually not a big deal.
1.) Unless you are computing slice time correction you can ignore this. So ignore this for DTI and ignore this for fMRI where you are not conducting slice time correction.
2.) In my experience this problem is either due to protocols that do not correctly populate the CSA header (a known issue with the excellent CMRR multi band sequences) or the well known problems Siemens has with not zeroing the CSA header. In either case, the problem might be fixed by using the latest version of dcm2niix which is aware of these bugs and will attempt to compensate for them.
3.) If the latest version of my software is not able to ferret out the correct slice order from your protocol you can get it by looking directly on your scanner console or on the protocol PDF you can generate on your scanner.
http://www.mccauslandcenter.sc.edu/crnl/tools/stc
http://www.mccauslandcenter.sc.edu/crnl/psyc589888
4.) As the cause of these issues has been identified, I suspect these problems will vanish at some point when you get a system or sequence upgrade.
1.) Unless you are computing slice time correction you can ignore this. So ignore this for DTI and ignore this for fMRI where you are not conducting slice time correction.
2.) In my experience this problem is either due to protocols that do not correctly populate the CSA header (a known issue with the excellent CMRR multi band sequences) or the well known problems Siemens has with not zeroing the CSA header. In either case, the problem might be fixed by using the latest version of dcm2niix which is aware of these bugs and will attempt to compensate for them.
3.) If the latest version of my software is not able to ferret out the correct slice order from your protocol you can get it by looking directly on your scanner console or on the protocol PDF you can generate on your scanner.
http://www.mccauslandcenter.sc.edu/crnl/tools/stc
http://www.mccauslandcenter.sc.edu/crnl/psyc589888
4.) As the cause of these issues has been identified, I suspect these problems will vanish at some point when you get a system or sequence upgrade.
Jul 6, 2016 07:07 AM | Loreen Mamerow
RE: Warning unknown Siemens slice order 41340006
Dear Chris,
That's great, thank you very much for your response!
Just to make sure I understand the implications of your answer: When I pull my images onto the dcm2nii surface, the output yields an interleaved acquisition, which is correct for our protocol. Therefore I take it that there is no actual problem with the converted images? Unless the conversion takes place based on an initially incorrect or unrecognized slice order?
Best wishes, Loreen
That's great, thank you very much for your response!
Just to make sure I understand the implications of your answer: When I pull my images onto the dcm2nii surface, the output yields an interleaved acquisition, which is correct for our protocol. Therefore I take it that there is no actual problem with the converted images? Unless the conversion takes place based on an initially incorrect or unrecognized slice order?
Best wishes, Loreen
Jul 6, 2016 12:07 PM | Chris Rorden
RE: Warning unknown Siemens slice order 41340006
Yes, if your dataset was interleaved simply specify this when
conducting your slice timing correction. All the warning is telling
you is that my software is not sure how to handle the slice_code
value stored in the NIfTI header. You can view this value created
by dcm2niix by running fslhd on any NIfTI dataset. Where possible
my software attempts to record the slice order in the header for
you so you do not forget, but this is not possible when the DICOM
CSA header is not correctly specified (e.g. in your case it seems
to be zero padded).
If dcm2niix was confident it could determine the slice order you will see
> fslhd fmri.nii
sizeof_hdr 348
data_type INT16
dim0 4
...
slice_name sequential_increasing
slice_code 1
if dcm2niix was not sure of the slice order you will see:
> fslhd fmri.nii
sizeof_hdr 348
data_type INT16
dim0 4
...
slice_name Unknown
slice_code 0
If dcm2niix was confident it could determine the slice order you will see
> fslhd fmri.nii
sizeof_hdr 348
data_type INT16
dim0 4
...
slice_name sequential_increasing
slice_code 1
if dcm2niix was not sure of the slice order you will see:
> fslhd fmri.nii
sizeof_hdr 348
data_type INT16
dim0 4
...
slice_name Unknown
slice_code 0