AFNI realtime Siemens SMS EPI error

AFNI Message Board
1

I am trying to implement a realtime fMRI neurofeedback application for the 7T Siemens (Terra.X) machine at my current institution. The scan sequence is a built-in Siemens simultaneous multi-slice (SMS) protocol. We are exporting the EPI data using a BOLD add-in feature (syngo MR XA30 platform) to our processing computer with AFNI (v24.1.06 April 26, 2024) running Ubuntu.

I am building off of the scripts that I successfully implemented in 2018 on the 3T Siemens Prisma at the NIH Clinical Center (with Vinai Roopchansingh). Everything seems to be working properly up to a point (AFNI opens in RT mode, computer receives DICOMS, Dimon (v.4.32) sends incoming DICOMS to AFNI). As far as I can tell, AFNI seems to be interpreting each volume (76 slices, TR = 1) as 76 individual volumes instead of slices. I have attached an image of the real time plot AFNI generates by default. The Dimon output below lists the volumes correctly, but lags (does not move to the next volume until AFNI finishes plotting the 76 slices).

afni_output_rtnf
afni_output_rtnf1870×1546 512 KB

Dimon output:
ART: comm link to afni established at
-- scanning for first volume
++Data detected to be oblique
**CID: have non-mosaic <full_file_path>.dcm with 76 images
*+ WARNING: Bad DICOM header - assuming oblique scaling direction!
**CID: have non-mosaic <full_file_path>.dcm with 76 images
no more warnings will be printed...

-- first volume found (1 slices)
-- scanning for additional volumes...
-- run ##: 1 2 3 4 5 6 7 8 9 10 11 12

Note that the script I run uses Dimon as follows:
Dimon -rt -quit -inf

I tried modifications to Dimon (below), and in non-real time mode, which produced the same behavior as described above.
"-num_slices 76"
"-assume_dicom_mosaic"

Is there something else I should implement so that AFNI parses the incoming data appropriately, or is this novel behavior? Thanks for your help.

-Samantha Fede

2

Hi Samantha,
Have you got it to work?
I am facing the same issue on XA60 but I don't know what BOLD Add-in is, I just export to a samba drive by setting ideacmdtool's parameter "sendIMA" to "ON".

I think that our problem is due to the fact that the DICOM are now enhanced and that Dimon cannot read enhanced DICOMs for the moment. See this thread: about Dimon and XA30.

So for offline data it seems to work with dcm2niix as it expects a directory as input but for a stream of DICOMs, it seems pretty difficult to use dcm2niix as it lacks the "-rt" option of Dimon.

Thank you,
Jonathan

3

There is also rumor of a compatibility option that would allow the scanners to write out the older form of DICOM images. It has not been verified as far as I know.

  • rick
4

We're still experiencing this issue. The compatibility option does not seem to be available for the rt export function-- we've looked and also asked our Siemens contacts, who said as far as he knows that mode is only designed to apply to the PACS export and cannot be applied to the real time export.

I sent a phantom DICOM to Rick a bit ago so hoping there's a solution the AFNI can develop? Thanks for your help, Rick.

5

Thank you Rick and Samantha,
I did a test using "interoperability" (i.e. not enhanced) and the real time exported DICOMs were still enhanced.

I am trying to unenhance the dicoms with help of gdcmtar to a temporary directory from where I could run Dimon (this time without option -assume_dicom_mosaic) but it might take some time to get it work since it generates Nslices times more files at once.
Jonathan

6

I'll be interested to hear if you can get something to work with this! Please let us know if you figure something out.