Question
Question Posted 03/05/25:
Hello, I have downloaded images for a cohort that I made using ARC builder. I have chosen file names ‘Axial_rsfMRI__Eyes_Open_’ and ‘Axial_rsfMRI__EYES_OPEN_’ for my resting-state scans, and the associated T1s I have found are named as follows: ‘Accelerated_Sagittal_MPRAGE’, ‘Accelerated_SAG_IR-SPGR’, ‘Accelerated_Sag_IR-FSPGR’, ‘Accelerated_Sagittal_IR-FSPGR’, ‘Sag_Accel_IR-FSPGR. How do I determine consistency of acquisition parameters? And should I have created an imaging collection rather than select the cohort and then download their imaging to select the scans of interest? Many thanks in advance for your help.
Hello, I have downloaded images for a cohort that I made using ARC builder. I have chosen file names ‘Axial_rsfMRI__Eyes_Open_’ and ‘Axial_rsfMRI__EYES_OPEN_’ for my resting-state scans, and the associated T1s I have found are named as follows: ‘Accelerated_Sagittal_MPRAGE’, ‘Accelerated_SAG_IR-SPGR’, ‘Accelerated_Sag_IR-FSPGR’, ‘Accelerated_Sagittal_IR-FSPGR’, ‘Sag_Accel_IR-FSPGR. How do I determine consistency of acquisition parameters? And should I have created an imaging collection rather than select the cohort and then download their imaging to select the scans of interest? Many thanks in advance for your help.
Response posted 03/05/25 by Jeff Gunter:
I would download all the images you may want and filter out those you don't. The DICOM tags for TR/TE/TI etc will be more reliable than SeriesDescriptions (which are almost certainly the source of the text strings in the question).
The IR-*SPGR sequences are GE; GE didn't license the MP-RAGE patent,; contrast is similar to MP-RAGE. I'll skip discussion of ADNI-1, since it had no fMRI in the protocol.
In ADNI-2/GO each subject had a fully sampled and an accelerated T1w sequence. In ADNI-3 the unaccelerated T1w images were dropped. In ADNI-4 the same T1w imaging was kept as in ADNI-3 and some newer more highly accelerated acquisitions added.
RS-fMRI was on Philips only for ADNI-2; in ADNI-3 and 4 it was/is on GE, Philips and Siemens. Broadly speaking there are TR=3sec and TR<1sec variants depending on scanner capabilities. (The Siemens scanners with support for simultaneous multi-slice imaging acquired data in the 600-800ms range for TR). ADNI-2 fMRI had a wall clock duration of about 10 minutes; in ADNI-3 that was cut to 5 minutes.
As in most multi center imaging studies, the DICOM SeriesDescription is hint as to the sequence. But it can generally be changed at the scanner console.
The IR-*SPGR sequences are GE; GE didn't license the MP-RAGE patent,; contrast is similar to MP-RAGE. I'll skip discussion of ADNI-1, since it had no fMRI in the protocol.
In ADNI-2/GO each subject had a fully sampled and an accelerated T1w sequence. In ADNI-3 the unaccelerated T1w images were dropped. In ADNI-4 the same T1w imaging was kept as in ADNI-3 and some newer more highly accelerated acquisitions added.
RS-fMRI was on Philips only for ADNI-2; in ADNI-3 and 4 it was/is on GE, Philips and Siemens. Broadly speaking there are TR=3sec and TR<1sec variants depending on scanner capabilities. (The Siemens scanners with support for simultaneous multi-slice imaging acquired data in the 600-800ms range for TR). ADNI-2 fMRI had a wall clock duration of about 10 minutes; in ADNI-3 that was cut to 5 minutes.
As in most multi center imaging studies, the DICOM SeriesDescription is hint as to the sequence. But it can generally be changed at the scanner console.
