Question

Question Posted 10/21/15:
Hello,
We are trying to do longitudinal study for brain and Hippocampus. I have searched in ADNI database and could find two different studies that provide brain mask ( "FreeSurfer Longitudinal Processing brainmask", "MIDAS Whole Brain Mask") but they are both scaled and registered to a template. I am not sure about the data named: "Total Intracranial Volume Brain Mask" because it has Spatial Normalization feature and I dont know if the coordinates are original or they normalized them as well. I also found study named "Hippocampal Mask" but this data is scaled to a template as well. Does anyone know of any brain mask and hippocampus mask that the data is not registered to any template and they have original coordination. I will be very appreciated if anyone can help me in this problem.

Thank you
Best Regards
Hajar
Response posted 10/27/15 by ADNI MRI Core:
Hello we received this information from the Lab at UCSF who was in charge of hippocampus measurements for ADNI. See below:


FreeSurfer brain mask (either from cross-sectional or longitudinal processing) is created in subject FreeSurfer image space. there shouldn't be any scaling or normalization. By default FreeSurfer conforms the input image to coronal orientation. this is captured by a transformation we uploaded to LONI-IDA. but even this transformation does not have any scaling or normalization factors. application of the inverse transformation will bring the FreeSurfer mask back to subject's native image space (this is well document at FreeSurfer wiki).

"Hippocampal Mask" is generated via SNT processing and it's defined in the subject's native image space.

"MIDAS Whole Brain Mask" is provided by Fox Lab part of their boundary shift integral method. Again, this is in the subject's native image space.

"MAPER Segmentation Mask" is provided by Heckemann Lab, and it is in the subject's native image space.

"Total Intracranial Volume Brain Mask" is estimated from PD/T2 images, not from T1 image. But it is in the subject's native image space.

Note that depending on the convention (radiological vs neurological) data might appear left-right flipped in different image visualization tools but all should be properly coded in the header files.



Hopefully this is helpful.


ADNI MRI
Response posted 10/27/15 by Hajar Hamidian:
Thanks alot for your respond and all the information. When I looked at Hippocampal mask data I can see MPR; GradWarp; B1 Correction; N3; Scaled for processing and when I check the xml file downloaded with the data I can see scale=1.002424,1.009383,1.01034. So I think the data is rigid registered to a template which I dont know what is it but I believe that the data is scaled. This can be seen in the data for longitudinal processing for brain mask as well. I have not checked all the other studies but I assumed that when the processing include Scaled, it means that the data is rigid registered to a template. Can you please correct me if I am wrong?

Thank you
Best Regards
Hajar