×
  • Select the area you would like to search.
  • ACTIVE INVESTIGATIONS Search for current projects using the investigator's name, institution, or keywords.
  • EXPERTS KNOWLEDGE BASE Enter keywords to search a list of questions and answers received and processed by the ADNI team.
  • ADNI PDFS Search any ADNI publication pdf by author, keyword, or PMID. Use an asterisk only to view all pdfs.

Question

Question Posted 09/19/20:
I've downloaded MRI data in the ANALYZE format, and I have noticed some inconsistencies in the file sizes. The GradWarp data are usually either of size 23.6 or 43.5MB. I can't figure out why these file sizes are different even though the pre-processing steps listed in the database are identical. This fact cuts my sample size in half because I need to use files that are all of the same dimension. Please help me understand these differences. Thank you!
Response posted 09/19/20 by Jeff Gunter:
It is likely that the larger files have been promoted to floating point during gradient distortion correction in post-processiing. Either recast the int16 data as floating point or round the floating point data and recast as 16 bit integers .

As an aside, you may have better success using NIfTI formatted data rather than Analyze formatted data.
Response posted 09/19/20 by Kristen Knight:
Thank you very much for your reply. I'm unfamiliar with this file format myself, but the MatLab program I'm using is easiest to run with the hdr/img files for each subject. (I have previously worked with the NiFTI images as well but they also have the various sizes for the files.) I did some more digging and found even more details for the images. I believe I have discovered that GE Medical has images of size 43.5MB and SIEMENS has size 23.6MB. The pixel spacing and number of pixels for the X,Y,Z axes are different for the two. Would recasting as 16 bit integers still fix this? It would be great to combine data from both scanners.

Another question, my MatLab program also asks for the slices in mm. For instance, in the SIEMENS 23.6MB data we have pixel sizes of Matrix X=192.0 pixels; Matrix Y=192.0 pixels; Matrix Z=160.0. Does this describe the size of the slices? If not, where may I find that info?
Response posted 09/19/20 by Jeff Gunter:
Unfortunately, the best answer I can give you is that you need to learn how to read the dimensions in pixels and the pixel spacing from the file headers.
If you are combining data with different voxel sizes for a program that requires all inputs have the same voxel dimensions and field of view then you have to resample data somewhere. That being said there is no reason to expect head position within the field of view to be consistent over time within subject nor across subjects. Since you haven't actually said what you are working on there's really not a good way to provide more useful suggestions.
Go back to list of topics >