Slicetiming warning on AP/PA sequences when using dcm2niix on Siemens XA30 data

1

Hello experts,

I have been running dcm2niix on a dataset, half of which was acquired before a Siemens upgrade and half afterwards (VE11 pre-upgrade/XA30 post-upgrade). Since the upgrade, I have been getting the following types of warnings about slice timing in my AP/PA files:

(sub-****_ses-01_dir-AP_epi)    stdout 2023-08-22T04:47:05.451382:Warning: Slice timing appears corrupted (range 0..13860, TR=7060 ms)
(sub-****_ses-01_dir-PA_epi)    stdout 2023-08-22T04:47:05.569298:Warning: Slice timing appears corrupted (range 0..13860, TR=7060 ms)
(sub-****_ses-01_magnitude)     WARNING: For now not embedding BIDS and info generated .nii.gz itself since sequence produced multiple files

Whereas before, the only warning was:

(sub-****_ses-01_magnitude)     WARNING: For now not embedding BIDS and info generated .nii.gz itself since sequence produced multiple files

(which I think has to do with 2 files generated for a multi-echo sequence).

From what I understand, Siemens doesn’t put the reliable slice timing for first image. Before upgrade, we had multiple files coming from this mapping sequence (1 image/slice, I think), now there is only one (1 image/volume, I think) So, now we can only extract slice timing from the unreliable first image.

However, I have checked the slice timing extracted in some json file both pre- and post-upgrade, and they have similar values and a similar range (Siemens does seem to introduce some jitter, enough to even change the order in some cases, but I think this is expected??). The range is about double the TR (which is the same pre/post upgrade). So, I’m not sure if anything has changed or if dcm2niix is throwing this warning simply because there is now only one file.

Does anyone have any insight into these warnings?

Thanks in advance!

[dcm2niix version: v1.0.20230411]

2

I suspect these warnings are spurious, as they are false alarms for some of the odd aspects of XA10 and XA11 that have been simplified with XA20 and XA30. If you are feeling bold, you can always compile the development branch, which has benefited from the widespread adoption of XA30:

git clone --branch development https://github.com/rordenlab/dcm2niix.git
cd dcm2niix/console
make
./dcm2niix
3

Thank you!

I don’t think we are bold enough for the development version at the moment, since we are using dcm2niix inside heudiconv with a docker image right now and it might be a little complicated (upgrading to the latest dcm2niix distribution was already a bit hacky). But it’s good to know that these warnings are probably not worth spending hours troubleshooting and we can move forward to the next issue. :slight_smile: