Hi,
Would it be a good idea to upsample the DWI images for HCP data? Given the data’s good quality and considering that my device has 8GB of RAM?
Thanks,
Yana
Hi,
Would it be a good idea to upsample the DWI images for HCP data? Given the data’s good quality and considering that my device has 8GB of RAM?
Thanks,
Yana
Hello Yana,
I cannot think of an application where upsampling of HCP data would be helpful. But just in case, could you tell me what are you going to do with the data?
(I assume you refer to the HCP data with 1.25 mm isometric voxel size for diffusion MRI)
Best,
Oren
Hi Oren,
Thanks for your reply. I would like to perform fixel-based analysis to compare between subjects fibre density derived from FODs. Yes I’m referring to that voxel size.
Thanks,
Yana
Hello Yana,
For fixel-basd analysis (FBA), it is pretty straightforward. If you check the current FBA documentation (https://mrtrix.readthedocs.io/en/latest/fixel_based_analysis/st_fibre_density_cross-section.html), you’ll see that step 7 recommends upsampling only if original images resolution is above 1.3 mm voxel size. HCP data have 1.25 mm voxel size, so no upsampling is required (notice that you also need to drop the -voxel_size 1.3 in the last command of step 10).
Other than that – notice that your device would not be able to run FBA analysis of data with 1.25 mm with an 1.25 mm template, as it doesn’t have enough memory. You’ll need 64GB+ for that (see discussion in thread Fixel Analysis Computer Question). The thread mentions, however, the possibility of downsampling the template to reduce memory requirements (if you don’t have access to a computer with a larger memory).
Good luck,
Oren
Hi Oren,
Thanks for your helpful reply.
Best,
Yana