I am new to MRS data acquisition and analysis, and I would appreciate any guidance on the following issue.
We have collected MRS data using the MEGA-PRESS sequence (downloaded from CRMM, VE111E) on a SIEMENS SKYRA 3T scanner. The dataset includes a VAPOR-ON sequence (approximately 6 minutes) and a VAPOR-OFF sequence (about 30 seconds). The raw data consists of four folders, as shown in the attached image: VAPOR-ON, VAPOR-ON-REF, VAPOR-OFF, and VAPOR-OFF-REF. Each folder contains two .IMA files.
An important observation is that all .IMA files are the same size, 128 KB. I have encountered difficulties transferring them using dcm2nii and loading them with Gannet.
My questions are as follows:
What are these .IMA files, and why is their size consistent across all of them?
How should I process these files? Is there a standard pipeline for handling this type of data?
Any insights or suggestions would be greatly appreciated!
May I ask how I can export the data in .rda or .dat format? Is there a way to convert these .ima files into .rda or .dat? I’m still unclear about the nature of these files—are they time series data, similar to BOLD signals?
Do you have any sort of documentation or manual for this sequence?
My guess would be that the VAPOR_ON and VAPOR_OFF files contain the water-suppressed MRS data (the edit-on and edit-off sub-experiments of MEGA-PRESS), and the respective _REF files are the water-unsuppressed files. But it isn’t very clear to me why each folder contains two files.
dcm2nii cannot handle DICOM that contain raw MRS data. You can try loading them with Osprey (see the documentation for how to set up the job file for DICOM MRS data), but without knowing the folder structure better it’s difficult for me to judge this from a distance. It is possible that you’ll need to put the IMA files from VAPOR_ON and VAPOR_OFF into a single folder, and the ones from VAPOR_ON_REF and VAPOR_OFF_REF into another.
