Duplicated image ID

[Mathieu Dubois]

Hello,

While inspecting the content of directories downloaded from IDA, I have noticed that some image ID are duplicated.

For instance, image ID 58872 corresponds to 2 files:
005_S_0572/MP-RAGE/2007-07-09_08_49_33.0/S34604/ADNI_005_S_0572_MR_MP-RAGE__br_raw_20070710090153930_1_S34604_I58872.nii
005_S_0572/MP-RAGE/2007-07-09_08_49_33.0/S34604/ADNI_005_S_0572_MR_MP-RAGE__br_raw_20070710090228153_167_S34604_I58872.nii
Those 2 images are close but not identical.

What is the difference between them and which image should I choose?

Mathieu

[Mathieu Dubois]

Hi,

Any news on this issue?

[Mike Donohue]

Hi Mathieu - 

You might try the MRI Core: http://adni.loni.usc.edu/support/experts-knowledge-base/?core=MRI+Core

Mike

[Mathieu Dubois]

Thanks Mike! I will contact them and keep the ML in touch.

[Mathieu Dubois]

Hello,

I received a response from MRI Core yesterday. If I understand their answer, there was an error in the protocol when acquiring the first image (too much slices) so there was a rescan. Therefore one should use the second file.

3 questions remain:
  1) I can't remember where to find the information about rescans.
  2) I don't know how to choose the correct image automatically: I could use some software to extract the number of slices from the images (assuming it's the only source of problems) or assume that the correct image is the one containing 167.
  3) I was thinking that the Image ID is assigned automatically when the image is entered in the DB (like in most DB) in which case this should not happen. Is the image ID assigned manually?

Mathieu

[Danielle J Harvey]

Hi Mathieu,

When I look in the MPRAGEMETA file, it seems as though the MRI Core chose SeriesID 34606 for this participant at the month 12 visit (3T scan) over 34604 (34606 got the GradWarp, B1 correction, N3). The two files you mention are from 34604. I don't recall if this question is related to finding raw images that correspond to a gradwarp/B1 corrected image or if it is something else.

I believe the Image IDs are assigned automatically, but I don't know that for certain.

I'm not sure if any of the information in the recall spreadsheet will be helpful, but you can find it at the following link (I believe most of these were just rescaling rather than rescanning)

http://adni.loni.usc.edu/updated-recall-speadsheet-for-mr-images-now-available-to-researchers/

Danielle

---

From: adni...@googlegroups.com <adni...@googlegroups.com> on behalf of 'Mathieu Dubois' via Alzheimer's Disease Neuroimaging Initiative (ADNI) Data <adni...@googlegroups.com>
Sent: Friday, June 5, 2015 9:32 AM
To: adni...@googlegroups.com
Subject: [adni-data] Re: Duplicated image ID

 

--
You received this message because you are subscribed to the Google Groups "Alzheimer's Disease Neuroimaging Initiative (ADNI) Data" group.
To unsubscribe from this group and stop receiving emails from it, send an email to adni-data+...@googlegroups.com.
To post to this group, send email to adni...@googlegroups.com.
Visit this group at http://groups.google.com/group/adni-data.
For more options, visit https://groups.google.com/d/optout.

[Mathieu Dubois]

Hi,

When I look in the MPRAGEMETA file, it seems as though the MRI Core chose SeriesID 34606 for this participant at the month 12 visit (3T scan) over 34604 (34606 got the GradWarp, B1 correction, N3). The two files you mention are from 34604. I don't recall if this question is related to finding raw images that correspond to a gradwarp/B1 corrected image or if it is something else.

You're right and this is the case for the 4 cases that I have found (image IDs 58872, 41449, 58423 and 70013). I haven't found all those images in MRIMPRANK but most of the time they are graded 1. Also, according to the tool I used (AimsFileInfo from BrainVisa), all those images have 166 slices.

This question is not related to finding raw images (I took a different strategy for this and I should close the question). I use MPRAGEMETA to identify original and fully pre-processed images which is what is inserted in my DB now. So the question of finding which is the correct original, while not crucial, is important.

 

I believe the Image IDs are assigned automatically, but I don't know that for certain.

OK. I guess those special cases where manually assigned the same ID.

 

I'm not sure if any of the information in the recall spreadsheet will be helpful, but you can find it at the following link (I believe most of these were just rescaling rather than rescanning)

http://adni.loni.usc.edu/updated-recall-speadsheet-for-mr-images-now-available-to-researchers/

 I haven't found any information about the images in question in this file.

[Mathieu Dubois]

Hi,

I have downloaded the DICOM images of those series (I generally use nifti). In all 4 cases, there are 332 DICOM images (labeled from _1 to _332): DICOM information says there are 332 slices but trying to convert to nifti with dcm2nii creates 2 files made of 166 slices (there's probably a field to says that but I don't know DICOM enough). It helped me to make sense of the MRI Core answer.

I guess I should just remove those images from my DB.