Question
Question Posted 02/01/25:
Dear Experts,
My name is Bence and i'm working with the ADNI3 database, especially the DTI images. I do all the postprocessing methods, and finally do a TBSS, segmented 50 white matter regions and compared the DTI metrics between 45-45 healthy and AD people. My biggest problem, that I won't able to do the Eddy current and distortion fixes, because I can't find the A-P P-A images.
My question is that I'm able to do the pre-processing without A-P and P-A images, and my research is totally wrong due to the deficiency of corrections method?
Thank for your answear,
Best Regard,
Bence
Dear Experts,
My name is Bence and i'm working with the ADNI3 database, especially the DTI images. I do all the postprocessing methods, and finally do a TBSS, segmented 50 white matter regions and compared the DTI metrics between 45-45 healthy and AD people. My biggest problem, that I won't able to do the Eddy current and distortion fixes, because I can't find the A-P P-A images.
My question is that I'm able to do the pre-processing without A-P and P-A images, and my research is totally wrong due to the deficiency of corrections method?
Thank for your answear,
Best Regard,
Bence
Response posted 02/01/25 by Rob Reid:
Hello Bence,
Most of ADNI3 does not have PA-AP pairs, because they triggered a bug in Siemens V series software that was not really resolved until ADNI4, and GE would have required changing a CV with the scanner (often clinical) in research mode. (Newer GE scanners have a better approach to the problem, which we use when possible in ADNI4.)
But you don't need AP-PA pairs to use eddy - just skip topup and proceed directly to eddy to correct eddy distortion and head motion. EPI distortion can be corrected in other ways, namely by calculating a warp to/from either an undistorted image (e.g. the T1w) or an image that may be distorted but is the same for everybody, such as a dMRI-based template. I believe TBSS falls into the latter category, but you should investigate its innards yourself to verify that.
Most of ADNI3 does not have PA-AP pairs, because they triggered a bug in Siemens V series software that was not really resolved until ADNI4, and GE would have required changing a CV with the scanner (often clinical) in research mode. (Newer GE scanners have a better approach to the problem, which we use when possible in ADNI4.)
But you don't need AP-PA pairs to use eddy - just skip topup and proceed directly to eddy to correct eddy distortion and head motion. EPI distortion can be corrected in other ways, namely by calculating a warp to/from either an undistorted image (e.g. the T1w) or an image that may be distorted but is the same for everybody, such as a dMRI-based template. I believe TBSS falls into the latter category, but you should investigate its innards yourself to verify that.
